Resolve timestep mismatch between dit forward and pred_noise_to_pred_video

checkpoint ode

checkpoint

fix t2v to i2v

lint

chekpt

checkpoint

hacked but working

ode_init scripts

WIP fixing time embedding

WIP fixing time embedding

checkpoint

update

fix

revert

revert

revert

update

update

visualize

.buildkite/pipeline.yml

@@ -176,3 +176,37 @@ steps:
                   - TEST_TYPE=precision_vsa
                 agents:
                   queue: "default"
+            - path:
+                - "csrc/attn/vmoba_attn/**"
+                - "pyproject.toml"
+                - "docker/Dockerfile.python3.12"
+              config:
+                command: "timeout 15m .buildkite/scripts/pr_test.sh"
+                label: "Precision Tests VMoBA"
+                env:
+                  - TEST_TYPE=precision_vmoba
+                agents:
+                  queue: "default"
+            - path:
+                - "csrc/attn/vmoba_attn/vmoba/**"
+                - "fastvideo/attention/backends/vmoba.py"
+                - "pyproject.toml"
+                - "docker/Dockerfile.python3.12"
+              config:
+                command: "timeout 15m .buildkite/scripts/pr_test.sh"
+                label: "Inference Tests VMoBA"
+                env:  
+                  - TEST_TYPE=inference_vmoba
+                agents:
+                  queue: "default"
+            - path:
+                - "fastvideo/**"
+                - "pyproject.toml"
+                - "docker/Dockerfile.python3.12"
+              config:
+                command: "timeout 15m .buildkite/scripts/pr_test.sh"
+                label: "Unit Tests"
+                env:
+                  - TEST_TYPE=unit_test
+                agents:
+                  queue: "default"

.buildkite/scripts/pr_test.sh

@@ -109,6 +109,19 @@ case "$TEST_TYPE" in
         log "Running distillation DMD tests..."
         MODAL_COMMAND="$MODAL_ENV WANDB_API_KEY=$WANDB_API_KEY python3 -m modal run $MODAL_TEST_FILE::run_distill_dmd_tests"
         ;;
+        # run_inference_tests_vmoba
+    "inference_vmoba")
+        log "Running V-MoBA inference tests..."
+        MODAL_COMMAND="$MODAL_ENV python3 -m modal run $MODAL_TEST_FILE::run_inference_tests_vmoba"
+        ;;
+    "precision_vmoba")
+        log "Running V-MoBA precision tests..."
+        MODAL_COMMAND="$MODAL_ENV python3 -m modal run $MODAL_TEST_FILE::run_precision_tests_vmoba"
+        ;;
+    "unit_test")
+        log "Running unit tests..."
+        MODAL_COMMAND="$MODAL_ENV python3 -m modal run $MODAL_TEST_FILE::run_unit_test"
+        ;;
     *)
         log "Error: Unknown test type: $TEST_TYPE"
         exit 1

.github/workflows/pr-test.yml

@@ -62,8 +62,8 @@ on:
         required: false
         default: false
         type: boolean
-      run_nightly_test:
-        description: "Run nightly-test"
+      run_unit_test:
+        description: "Run unit-test"
         required: false
         default: false
         type: boolean
@@ -93,6 +93,7 @@ jobs:
       inference-test-STA: ${{ steps.filter.outputs.inference-test-STA }}
       precision-test-STA: ${{ steps.filter.outputs.precision-test-STA }}
       precision-test-VSA: ${{ steps.filter.outputs.precision-test-VSA }}
+      unit-test: ${{ steps.filter.outputs.unit-test }}
     steps:
       - uses: actions/checkout@v4
       - uses: dorny/paths-filter@v3
@@ -102,6 +103,8 @@ jobs:
             # Define reusable path patterns
             common-paths: &common-paths
               - 'pyproject.toml'
+              - 'docker/Dockerfile.python3.10'
+              - 'docker/Dockerfile.python3.11'
               - 'docker/Dockerfile.python3.12'
             sta-kernel-paths: &sta-kernel-paths
               - 'csrc/attn/sliding_tile_attn/**'
@@ -155,6 +158,9 @@ jobs:
             precision-test-VSA:
               - *common-paths
               - *vsa-kernel-paths
+            unit-test:
+              - 'fastvideo/**'
+              - *common-paths
 
   encoder-test:
     needs: change-filter
@@ -235,7 +241,7 @@ jobs:
     secrets:
       RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
       RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
-  
+
   training-test:
     needs: change-filter
     if: >-
@@ -333,23 +339,42 @@ jobs:
       RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
       RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
 
-  nightly-test:
+  unit-test:
+    needs: change-filter
     if: >-
-      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_nightly_test == 'true')
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.unit-test == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_unit_test == 'true')
     uses: ./.github/workflows/runpod-test.yml
     with:
-      job_id: "nightly-test"
-      gpu_type: "NVIDIA A40"
-      gpu_count: 4
+      job_id: "unit-test"
+      gpu_type: "NVIDIA L40S"
+      gpu_count: 1
       volume_size: 100
       disk_size: 100
       image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
-      test_command: "wandb login $WANDB_API_KEY && uv pip install -e .[test] && pytest ./fastvideo/tests/nightly/test_e2e_overfit_single_sample.py -vs"
+      test_command: "uv pip install -e .[test] && pytest ./fastvideo/dataset/ -vs && pytest ./fastvideo/workflow/ -vs"
       timeout_minutes: 30
     secrets:
       RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
       RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
-      WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
+
+  # nightly-test:
+  #   if: >-
+  #     (github.event_name == 'workflow_dispatch' && github.event.inputs.run_nightly_test == 'true')
+  #   uses: ./.github/workflows/runpod-test.yml
+  #   with:
+  #     job_id: "nightly-test"
+  #     gpu_type: "NVIDIA A40"
+  #     gpu_count: 4
+  #     volume_size: 100
+  #     disk_size: 100
+  #     image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+  #     test_command: "wandb login $WANDB_API_KEY && uv pip install -e .[test] && pytest ./fastvideo/tests/nightly/test_e2e_overfit_single_sample.py -vs"
+  #     timeout_minutes: 30
+  #   secrets:
+  #     RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+  #     RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+  #     WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
 
   runpod-cleanup:
     # Add other jobs to this list as you create them
@@ -373,4 +398,4 @@ jobs:
           JOB_IDS: '["encoder-test", "vae-test", "transformer-test", "ssim-test-py3.10", "ssim-test-py3.11", "ssim-test-py3.12", "training-test", "training-test-VSA", "inference-test-STA", "precision-test-STA", "precision-test-VSA"]'
           RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
           GITHUB_RUN_ID: ${{ github.run_id }}
-        run: python .github/scripts/runpod_cleanup.py
+        run: python .github/scripts/runpod_cleanup.py

.gitignore

@@ -64,3 +64,6 @@ docs/source/distillation/examples/
 !docs/source/_static/images/**/*.png
 !comfyui/assets/**/*.png
 !comfyui/assets/**/*.gif
+
+dmd_t2v_output/
+preprocess_output_text/

README.md

@@ -7,7 +7,7 @@
 FastVideo features an end-to-end unified pipeline for accelerating diffusion models, starting from data preprocessing to model training, finetuning, distillation, and inference. FastVideo is designed to be modular and extensible, allowing users to easily add new optimizations and techniques. Whether it is training-free optimizations or post-training optimizations, FastVideo has you covered.
 
 <p align="center">
-    | 🕹️ <a href="https://fastwan.fastvideo.org/"<b>Online Demo</b></a> | <a href="https://hao-ai-lab.github.io/FastVideo"><b>Documentation</b></a> | <a href="https://hao-ai-lab.github.io/FastVideo/inference/inference_quick_start.html"><b> Quick Start</b></a> | 🤗 <a href="https://huggingface.co/collections/FastVideo/fastwan-6886a305d9799c8cd1496408"  target="_blank"><b>FastWan</b></a>  | 🟣💬 <a href="https://join.slack.com/t/fastvideo/shared_invite/zt-38u6p1jqe-yDI1QJOCEnbtkLoaI5bjZQ" target="_blank"> <b>Slack</b> </a> |  🟣💬 <a href="https://ibb.co/rG0QpZdw" target="_blank"> <b> WeChat </b> </a> |
+    | 🕹️ <a href="https://fastwan.fastvideo.org/"<b>Online Demo</b></a> | <a href="https://hao-ai-lab.github.io/FastVideo"><b>Documentation</b></a> | <a href="https://hao-ai-lab.github.io/FastVideo/inference/inference_quick_start.html"><b> Quick Start</b></a> | 🤗 <a href="https://huggingface.co/collections/FastVideo/fastwan-6886a305d9799c8cd1496408"  target="_blank"><b>FastWan</b></a>  | 🟣💬 <a href="https://join.slack.com/t/fastvideo/shared_invite/zt-3csdw1isz-Euq8_Q8~baewG8hxjXs2gQ" target="_blank"> <b>Slack</b> </a> |  🟣💬 <a href="https://ibb.co/S7HLCSTh" target="_blank"> <b> WeChat </b> </a> |
 </p>
 
 <div align="center">

csrc/attn/vmoba_attn/README.md

@@ -0,0 +1,32 @@
+# Attention Kernel Used in FastVideo
+
+## VMoBA: Mixture-of-Block Attention for Video Diffusion Models (VMoBA)
+
+### Installation
+Please ensure that you have installed FlashAttention version **2.7.1 or higher**, as some interfaces have changed in recent releases.
+
+### Usage
+
+You can use `moba_attn_varlen` in the following ways:
+
+**Install from source:**
+```bash
+python setup.py install
+```
+
+**Import after installation:**
+```python
+from vmoba import moba_attn_varlen
+```
+
+**Or import directly from the project root:**
+```python
+from csrc.attn.vmoba_attn.vmoba import moba_attn_varlen
+```
+
+### Verify if you have successfully installed
+
+```bash
+python csrc/attn/vmoba_attn/vmoba/vmoba.py
+```
+

csrc/attn/vmoba_attn/setup.py

@@ -0,0 +1,24 @@
+# SPDX-License-Identifier: Apache-2.0
+
+from setuptools import find_packages, setup
+
+PACKAGE_NAME = "vmoba"
+VERSION = "0.0.0"
+AUTHOR = "JianzongWu"
+DESCRIPTION = "VMoBA: Mixture-of-Block Attention for Video Diffusion Models"
+URL = "https://github.com/KwaiVGI/VMoBA"
+
+setup(
+    name=PACKAGE_NAME,
+    version=VERSION,
+    author=AUTHOR,
+    description=DESCRIPTION,
+    url=URL,
+    packages=find_packages(),
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: Apache Software License",
+    ],
+    python_requires='>=3.12',
+    install_requires=[]
+)

csrc/attn/vmoba_attn/tests/test_vmoba_attn.py

@@ -0,0 +1,97 @@
+# SPDX-License-Identifier: Apache-2.0
+
+import torch
+import pytest
+import random
+from csrc.attn.vmoba_attn.vmoba  import moba_attn_varlen
+
+def generate_test_data(batch_size, total_seqlen, num_heads, head_dim, dtype, device="cuda"):
+    """
+    Generates random data for testing the variable-length attention function.
+    """
+    torch.manual_seed(42)
+    random.seed(42)
+    torch.cuda.manual_seed_all(42)
+
+    # Generate sequence lengths for each item in the batch
+    if batch_size > 1:
+        # Ensure sequence lengths are reasonably distributed
+        avg_seqlen = total_seqlen // batch_size
+        seqlens = [random.randint(avg_seqlen // 2, avg_seqlen + avg_seqlen // 2) for _ in range(batch_size - 1)]
+        remaining_len = total_seqlen - sum(seqlens)
+        if remaining_len > 0:
+            seqlens.append(remaining_len)
+        else: # Adjust if sum exceeds total_seqlen
+            seqlens.append(avg_seqlen)
+            current_sum = sum(seqlens)
+            seqlens[-1] -= (current_sum - total_seqlen)
+        # Ensure all lengths are positive
+        seqlens = [max(1, s) for s in seqlens]
+        # Final adjustment to match total_seqlen
+        seqlens[-1] += total_seqlen - sum(seqlens)
+
+    else:
+        seqlens = [total_seqlen]
+
+    cu_seqlens = torch.tensor([0] + list(torch.cumsum(torch.tensor(seqlens), 0)), device=device, dtype=torch.int32)
+    max_seqlen = max(seqlens) if seqlens else 0
+
+    q = torch.randn((total_seqlen, num_heads, head_dim), dtype=dtype, device=device, requires_grad=False)
+    k = torch.randn((total_seqlen, num_heads, head_dim), dtype=dtype, device=device, requires_grad=False)
+    v = torch.randn((total_seqlen, num_heads, head_dim), dtype=dtype, device=device, requires_grad=False)
+
+    return q, k, v, cu_seqlens, max_seqlen
+
+
+@pytest.mark.parametrize("batch_size", [1, 2])
+@pytest.mark.parametrize("total_seqlen", [512, 1024])
+@pytest.mark.parametrize("num_heads", [8])
+@pytest.mark.parametrize("head_dim", [64])
+@pytest.mark.parametrize("moba_chunk_size", [64])
+@pytest.mark.parametrize("moba_topk", [2, 4])
+@pytest.mark.parametrize("select_mode", ["topk", "threshold"])
+@pytest.mark.parametrize("threshold_type", ["query_head", "head_global", "overall"])
+@pytest.mark.parametrize("dtype", [torch.float32, torch.float16, torch.bfloat16])
+def test_moba_attn_varlen_forward(
+    batch_size, total_seqlen, num_heads, head_dim, moba_chunk_size, moba_topk, select_mode, threshold_type, dtype
+):
+    """
+    Tests the forward pass of moba_attn_varlen for basic correctness.
+    It checks output shape, dtype, and for the presence of NaNs/Infs.
+    """
+    if dtype == torch.float32:
+        pytest.skip("float32 is not supported in flash attention")
+
+    q, k, v, cu_seqlens, max_seqlen = generate_test_data(
+        batch_size, total_seqlen, num_heads, head_dim, dtype
+    )
+
+    # Ensure chunk size is not larger than the smallest sequence length
+    min_seqlen = (cu_seqlens[1:] - cu_seqlens[:-1]).min().item()
+    if moba_chunk_size > min_seqlen:
+        pytest.skip("moba_chunk_size is larger than the minimum sequence length in the batch")
+
+    try:
+        output = moba_attn_varlen(
+            q=q,
+            k=k,
+            v=v,
+            cu_seqlens=cu_seqlens,
+            max_seqlen=max_seqlen,
+            moba_chunk_size=moba_chunk_size,
+            moba_topk=moba_topk,
+            select_mode=select_mode,
+            threshold_type=threshold_type,
+            simsum_threshold=0.5, # A reasonable default for threshold mode
+        )
+    except Exception as e:
+        pytest.fail(f"moba_attn_varlen forward pass failed with exception: {e}")
+
+    # 1. Check output shape
+    assert output.shape == q.shape, f"Expected output shape {q.shape}, but got {output.shape}"
+
+    # 2. Check output dtype
+    assert output.dtype == q.dtype, f"Expected output dtype {q.dtype}, but got {output.dtype}"
+
+    # 3. Check for NaNs or Infs in the output
+    assert torch.all(torch.isfinite(output)), "Output contains NaN or Inf values"

csrc/attn/vmoba_attn/vmoba/__init__.py

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: Apache-2.0
+from .vmoba import moba_attn_varlen, process_moba_input, process_moba_output

csrc/attn/vmoba_attn/vmoba/vmoba.py

@@ -0,0 +1,860 @@
+# SPDX-License-Identifier: Apache-2.0
+# Adapt from https://github.com/KwaiVGI/VMoBA/blob/main/src/vmoba.py
+
+import random
+import time
+import os
+import torch
+from typing import Tuple
+from flash_attn import flash_attn_varlen_func  # Use the new flash attention function
+from flash_attn.flash_attn_interface import _flash_attn_varlen_forward, _flash_attn_varlen_backward
+from functools import lru_cache
+from einops import rearrange
+
+
+@lru_cache(maxsize=16)
+def calc_chunks(cu_seqlen, moba_chunk_size):
+    """
+    Calculate chunk boundaries.
+
+    For vision tasks we include all chunks (even the last one which might be shorter)
+    so that every chunk can be selected.
+    """
+    batch_sizes = cu_seqlen[1:] - cu_seqlen[:-1]
+    batch_num_chunk = (batch_sizes + (moba_chunk_size - 1)) // moba_chunk_size
+    cu_num_chunk = torch.ones(
+        batch_num_chunk.numel() + 1,
+        device=cu_seqlen.device,
+        dtype=batch_num_chunk.dtype,
+    )
+    cu_num_chunk[1:] = batch_num_chunk.cumsum(dim=0)
+    num_chunk = cu_num_chunk[-1]
+    chunk_sizes = torch.full(
+        (num_chunk + 1,), moba_chunk_size, dtype=torch.int32, device=cu_seqlen.device
+    )
+    chunk_sizes[0] = 0
+    batch_last_chunk_size = batch_sizes - (batch_num_chunk - 1) * moba_chunk_size
+    chunk_sizes[cu_num_chunk[1:]] = batch_last_chunk_size
+    cu_chunk = chunk_sizes.cumsum(dim=-1, dtype=torch.int32)
+    chunk_to_batch = torch.zeros(
+        (num_chunk,), dtype=torch.int32, device=cu_seqlen.device
+    )
+    chunk_to_batch[cu_num_chunk[1:-1]] = 1
+    chunk_to_batch = chunk_to_batch.cumsum(dim=0, dtype=torch.int32)
+
+    # Do not filter out any chunk
+    filtered_chunk_indices = torch.arange(
+        num_chunk, device=cu_seqlen.device, dtype=torch.int32
+    )
+    num_filtered_chunk = num_chunk
+
+    return cu_chunk, filtered_chunk_indices, num_filtered_chunk, chunk_to_batch
+
+
+# --- Threshold Selection Helper Functions ---
+
+def _select_threshold_query_head(
+    gate: torch.Tensor,
+    valid_gate_mask: torch.Tensor,
+    gate_self_chunk_mask: torch.Tensor,
+    simsum_threshold: float
+) -> torch.Tensor:
+    """
+    Selects chunks for each <query, head> pair based on threshold.
+    Normalization and sorting happen along the chunk dimension (dim=0).
+    """
+    C, H, S = gate.shape
+    eps = 1e-6
+
+    # LSE‐style normalization per <head, query> (across chunks)
+    gate_masked = torch.where(valid_gate_mask, gate, -torch.inf) # Use -inf for max
+    gate_min_val = torch.where(valid_gate_mask, gate, torch.inf) # Use +inf for min
+
+    row_min = gate_min_val.amin(dim=0)        # (H, S)
+    row_max = gate_masked.amax(dim=0)         # (H, S)
+    denom = row_max - row_min
+    denom = torch.where(denom <= eps, torch.ones_like(denom), denom) # avoid divide‑by‑zero
+
+    gate_norm = (gate - row_min.unsqueeze(0)) / denom.unsqueeze(0)
+    gate_norm = torch.where(valid_gate_mask, gate_norm, 0.0) # (C, H, S)
+
+    # 1) pull out the self‐chunk’s normalized weight for each <head,seq>
+    self_norm = (gate_norm * gate_self_chunk_mask).sum(dim=0)  # (H, S)
+
+    # 2) compute how much more normalized weight we need beyond self
+    total_norm_sum = gate_norm.sum(dim=0) # (H, S)
+    remain_ratio = simsum_threshold - self_norm / (total_norm_sum + eps)  # (H, S)
+    remain_ratio = torch.clamp(remain_ratio, min=0.0) # if already ≥ thresh, no extra needed
+
+    # 3) zero out the self‐chunk in a copy, so we only sort “others”
+    others_norm = gate_norm.clone()
+    others_norm[gate_self_chunk_mask] = 0.0
+
+    # 4) sort the other chunks by descending norm, per <head,seq>
+    sorted_norm, sorted_idx = torch.sort(others_norm, descending=True, dim=0) # (C, H, S)
+
+    # 5) cumulative‑sum the sorted norms per <head,seq>
+    cumsum_others = sorted_norm.cumsum(dim=0) # (C, H, S)
+
+    # 6) for each <head,seq>, find the smallest k where cumsum_ratio ≥ remain_ratio
+    ratio = cumsum_others / (total_norm_sum.unsqueeze(0) + eps) # (C, H, S)
+    cond = ratio >= remain_ratio.unsqueeze(0) # (C, H, S) boolean mask
+    any_cond = cond.any(dim=0) # (H, S)
+    # Find the index of the first True value along dim 0. If none, use C-1.
+    cutoff = torch.where(any_cond, cond.float().argmax(dim=0), torch.full_like(any_cond, fill_value=C - 1)) # (H, S)
+
+    # 7) build a mask in sorted order up to that cutoff
+    idx_range = torch.arange(C, device=gate.device).view(-1, 1, 1) # (C, 1, 1)
+    sorted_mask = idx_range <= cutoff.unsqueeze(0) # (C, H, S)
+
+    # 8) scatter it back to original chunk order
+    others_mask = torch.zeros_like(gate, dtype=torch.bool)
+    others_mask.scatter_(0, sorted_idx, sorted_mask)
+
+    # 9) finally, include every self‐chunk plus all selected others
+    final_gate_mask = valid_gate_mask & (others_mask | gate_self_chunk_mask)
+
+    return final_gate_mask
+
+
+def _select_threshold_block(
+    gate: torch.Tensor,
+    valid_gate_mask: torch.Tensor,
+    gate_self_chunk_mask: torch.Tensor,
+    simsum_threshold: float
+) -> torch.Tensor:
+    """
+    Selects <query, head> pairs for each block based on threshold.
+    Normalization and sorting happen across the head and sequence dimensions (dim=1, 2).
+    """
+    C, H, S = gate.shape
+    HS = H * S
+    eps = 1e-6
+
+    # LSE‐style normalization per block (across heads and queries)
+    gate_masked = torch.where(valid_gate_mask, gate, -torch.inf) # Use -inf for max
+    gate_min_val = torch.where(valid_gate_mask, gate, torch.inf) # Use +inf for min
+
+    block_max = gate_masked.amax(dim=(1, 2), keepdim=True) # (C, 1, 1)
+    block_min = gate_min_val.amin(dim=(1, 2), keepdim=True) # (C, 1, 1)
+    block_denom = block_max - block_min
+    block_denom = torch.where(block_denom <= eps, torch.ones_like(block_denom), block_denom) # (C, 1, 1)
+
+    gate_norm = (gate - block_min) / block_denom # (C, H, S)
+    gate_norm = torch.where(valid_gate_mask, gate_norm, 0.0) # (C, H, S)
+
+    # 1) identify normalized weights of entries that *are* self-chunks (from query perspective)
+    self_norm_entries = gate_norm * gate_self_chunk_mask # (C, H, S)
+    # Sum these weights *per block*
+    self_norm_sum_per_block = self_norm_entries.sum(dim=(1, 2)) # (C,)
+
+    # 2) compute how much more normalized weight each block needs beyond its self-chunk contributions
+    total_norm_sum_per_block = gate_norm.sum(dim=(1, 2)) # (C,)
+    remain_ratio = simsum_threshold - self_norm_sum_per_block / (total_norm_sum_per_block + eps) # (C,)
+    remain_ratio = torch.clamp(remain_ratio, min=0.0) # (C,)
+
+    # 3) zero out the self‐chunk entries in a copy, so we only sort “others”
+    others_norm = gate_norm.clone()
+    others_norm[gate_self_chunk_mask] = 0.0 # Zero out self entries
+
+    # 4) sort the other <head, seq> pairs by descending norm, per block
+    others_flat = others_norm.contiguous().view(C, HS) # (C, H*S)
+    sorted_others_flat, sorted_indices_flat = torch.sort(others_flat, dim=1, descending=True) # (C, H*S)
+
+    # 5) cumulative‑sum the sorted norms per block
+    cumsum_others_flat = sorted_others_flat.cumsum(dim=1) # (C, H*S)
+
+    # 6) for each block, find the smallest k where cumsum_ratio ≥ remain_ratio
+    ratio_flat = cumsum_others_flat / (total_norm_sum_per_block.unsqueeze(1) + eps) # (C, H*S)
+    cond_flat = ratio_flat >= remain_ratio.unsqueeze(1) # (C, H*S) boolean mask
+    any_cond = cond_flat.any(dim=1) # (C,)
+    # Find the index of the first True value along dim 1. If none, use HS-1.
+    cutoff_flat = torch.where(any_cond, cond_flat.float().argmax(dim=1), torch.full_like(any_cond, fill_value=HS - 1)) # (C,)
+
+    # 7) build a mask in sorted order up to that cutoff per block
+    idx_range_flat = torch.arange(HS, device=gate.device).unsqueeze(0) # (1, H*S)
+    sorted_mask_flat = idx_range_flat <= cutoff_flat.unsqueeze(1) # (C, H*S)
+
+    # 8) scatter it back to original <head, seq> order per block
+    others_mask_flat = torch.zeros_like(others_flat, dtype=torch.bool) # (C, H*S)
+    others_mask_flat.scatter_(1, sorted_indices_flat, sorted_mask_flat)
+    others_mask = others_mask_flat.view(C, H, S) # (C, H, S)
+
+    # 9) finally, include every self‐chunk entry plus all selected others
+    final_gate_mask = valid_gate_mask & (others_mask | gate_self_chunk_mask)
+
+    return final_gate_mask
+
+
+def _select_threshold_overall(
+    gate: torch.Tensor,
+    valid_gate_mask: torch.Tensor,
+    gate_self_chunk_mask: torch.Tensor,
+    simsum_threshold: float
+) -> torch.Tensor:
+    """
+    Selects <chunk, query, head> triplets globally based on threshold.
+    Normalization and sorting happen across all valid entries.
+    """
+    C, H, S = gate.shape
+    CHS = C * H * S
+    eps = 1e-6
+
+    # LSE‐style normalization globally across all valid entries
+    gate_masked = torch.where(valid_gate_mask, gate, -torch.inf) # Use -inf for max
+    gate_min_val = torch.where(valid_gate_mask, gate, torch.inf) # Use +inf for min
+
+    overall_max = gate_masked.max() # scalar
+    overall_min = gate_min_val.min() # scalar
+    overall_denom = overall_max - overall_min
+    overall_denom = torch.where(overall_denom <= eps, torch.tensor(1.0, device=gate.device, dtype=gate.dtype), overall_denom)
+
+    gate_norm = (gate - overall_min) / overall_denom # (C, H, S)
+    gate_norm = torch.where(valid_gate_mask, gate_norm, 0.0) # (C, H, S)
+
+    # 1) identify normalized weights of entries that *are* self-chunks
+    self_norm_entries = gate_norm * gate_self_chunk_mask # (C, H, S)
+    # Sum these weights globally
+    self_norm_sum_overall = self_norm_entries.sum() # scalar
+
+    # 2) compute how much more normalized weight is needed globally beyond self-chunk contributions
+    total_norm_sum_overall = gate_norm.sum() # scalar
+    remain_ratio = simsum_threshold - self_norm_sum_overall / (total_norm_sum_overall + eps) # scalar
+    remain_ratio = torch.clamp(remain_ratio, min=0.0) # scalar
+
+    # 3) zero out the self‐chunk entries in a copy, so we only sort “others”
+    others_norm = gate_norm.clone()
+    others_norm[gate_self_chunk_mask] = 0.0 # Zero out self entries
+
+    # 4) sort all other entries by descending norm, globally
+    others_flat = others_norm.flatten() # (C*H*S,)
+    valid_others_mask_flat = valid_gate_mask.flatten() & ~gate_self_chunk_mask.flatten() # Mask for valid, non-self entries
+    
+    # Only sort the valid 'other' entries
+    valid_others_indices = torch.where(valid_others_mask_flat)[0]
+    valid_others_values = others_flat[valid_others_indices]
+
+    sorted_others_values, sort_perm = torch.sort(valid_others_values, descending=True) # (N_valid_others,)
+    sorted_original_indices = valid_others_indices[sort_perm] # Original indices in C*H*S space, sorted by value
+
+    # 5) cumulative‑sum the sorted valid 'other' norms globally
+    cumsum_others_values = sorted_others_values.cumsum(dim=0) # (N_valid_others,)
+
+    # 6) find the smallest k where cumsum_ratio ≥ remain_ratio globally
+    ratio_values = cumsum_others_values / (total_norm_sum_overall + eps) # (N_valid_others,)
+    cond_values = ratio_values >= remain_ratio # (N_valid_others,) boolean mask
+    any_cond = cond_values.any() # scalar
+    
+    # Find the index of the first True value in the *sorted* list. If none, use all valid others.
+    cutoff_idx_in_sorted = torch.where(
+        any_cond, 
+        cond_values.float().argmax(dim=0), 
+        torch.tensor(len(sorted_others_values) - 1, device=gate.device, dtype=torch.long)
+    )
+
+    # 7) build a mask selecting the top-k others based on the cutoff
+    # Select the original indices corresponding to the top entries in the sorted list
+    selected_other_indices = sorted_original_indices[:cutoff_idx_in_sorted + 1]
+
+    # 8) create the mask in the original flat shape
+    others_mask_flat = torch.zeros_like(others_flat, dtype=torch.bool) # (C*H*S,)
+    if selected_other_indices.numel() > 0: # Check if any 'other' indices were selected
+        others_mask_flat[selected_other_indices] = True
+    others_mask = others_mask_flat.view(C, H, S) # (C, H, S)
+
+    # 9) finally, include every self‐chunk entry plus all selected others
+    final_gate_mask = valid_gate_mask & (others_mask | gate_self_chunk_mask)
+
+    return final_gate_mask
+
+
+def _select_threshold_head_global(
+    gate: torch.Tensor,
+    valid_gate_mask: torch.Tensor,
+    gate_self_chunk_mask: torch.Tensor,
+    simsum_threshold: float
+) -> torch.Tensor:
+    """
+    Selects <chunk, query> globally for each head based on threshold.
+    """
+    C, H, S = gate.shape
+    eps = 1e-6
+
+    # 1) LSE‐style normalization per head (across chunks and sequence dims)
+    gate_masked = torch.where(valid_gate_mask, gate, -torch.inf)
+    gate_min_val = torch.where(valid_gate_mask, gate, torch.inf)
+
+    max_per_head = gate_masked.amax(dim=(0, 2), keepdim=True)  # (1, H, 1)
+    min_per_head = gate_min_val.amin(dim=(0, 2), keepdim=True)  # (1, H, 1)
+    denom = max_per_head - min_per_head
+    denom = torch.where(denom <= eps, torch.ones_like(denom), denom)
+
+    gate_norm = (gate - min_per_head) / denom
+    gate_norm = torch.where(valid_gate_mask, gate_norm, 0.0)  # (C, H, S)
+
+    # 2) sum normalized self‐chunk contributions per head
+    self_norm_sum = (gate_norm * gate_self_chunk_mask).sum(dim=(0, 2))  # (H,)
+
+    # 3) total normalized sum per head
+    total_norm_sum = gate_norm.sum(dim=(0, 2))  # (H,)
+
+    # 4) how much more normalized weight needed per head
+    remain_ratio = simsum_threshold - self_norm_sum / (total_norm_sum + eps)  # (H,)
+    remain_ratio = torch.clamp(remain_ratio, min=0.0)
+
+    # 5) zero out self‐chunk entries to focus on "others"
+    others_norm = gate_norm.clone()
+    others_norm[gate_self_chunk_mask] = 0.0  # (C, H, S)
+
+    # 6) flatten chunk and sequence dims, per head
+    CS = C * S
+    others_flat = others_norm.permute(1, 0, 2).reshape(H, CS)  # (H, C*S)
+    valid_flat  = (valid_gate_mask & ~gate_self_chunk_mask) \
+                   .permute(1, 0, 2).reshape(H, CS)           # (H, C*S)
+
+    # 7) vectorized selection of “others” per head
+    masked_flat = torch.where(valid_flat, others_flat, torch.zeros_like(others_flat))
+    sorted_vals, sorted_idx = torch.sort(masked_flat, dim=1, descending=True)  # (H, C*S)
+
+    cumsum_vals = sorted_vals.cumsum(dim=1)                                    # (H, C*S)
+    ratio_vals  = cumsum_vals / (total_norm_sum.unsqueeze(1) + eps)            # (H, C*S)
+    cond        = ratio_vals >= remain_ratio.unsqueeze(1)                      # (H, C*S)
+
+    has_cutoff = cond.any(dim=1)                                               # (H,)
+    default    = torch.full((H,), CS - 1, device=gate.device, dtype=torch.long)
+    cutoff     = torch.where(has_cutoff, cond.float().argmax(dim=1), default)  # (H,)
+
+    idx_range   = torch.arange(CS, device=gate.device).unsqueeze(0)            # (1, C*S)
+    sorted_mask = idx_range <= cutoff.unsqueeze(1)                             # (H, C*S)
+
+    selected_flat = torch.zeros_like(valid_flat)                               # (H, C*S)
+    selected_flat.scatter_(1, sorted_idx, sorted_mask)                         # (H, C*S)
+
+    # 8) reshape selection mask back to (C, H, S)
+    others_mask = selected_flat.reshape(H, C, S).permute(1, 0, 2)               # (C, H, S)
+
+    # 9) include self‐chunks plus selected others, and obey valid mask
+    final_gate_mask = valid_gate_mask & (gate_self_chunk_mask | others_mask)
+
+    return final_gate_mask
+
+
+class MixedAttention(torch.autograd.Function):
+    @staticmethod
+    def forward(
+        ctx,
+        q,
+        k,
+        v,
+        self_attn_cu_seqlen,
+        moba_q,
+        moba_kv,
+        moba_cu_seqlen_q,
+        moba_cu_seqlen_kv,
+        max_seqlen,
+        moba_chunk_size,
+        moba_q_sh_indices,
+    ):
+        ctx.max_seqlen = max_seqlen
+        ctx.moba_chunk_size = moba_chunk_size
+        ctx.softmax_scale = softmax_scale = q.shape[-1] ** (-0.5)
+
+        # Non-causal self-attention branch
+        # return out, softmax_lse, S_dmask, rng_state
+        self_attn_out_sh, self_attn_lse_hs, _, _ = _flash_attn_varlen_forward(
+            q=q,
+            k=k,
+            v=v,
+            cu_seqlens_q=self_attn_cu_seqlen,
+            cu_seqlens_k=self_attn_cu_seqlen,
+            max_seqlen_q=max_seqlen,
+            max_seqlen_k=max_seqlen,
+            softmax_scale=softmax_scale,
+            causal=False,
+            dropout_p=0.0,
+        )
+        # MOBA attention branch (non-causal)
+        moba_attn_out, moba_attn_lse_hs, _, _ = _flash_attn_varlen_forward(
+            q=moba_q,
+            k=moba_kv[:, 0],
+            v=moba_kv[:, 1],
+            cu_seqlens_q=moba_cu_seqlen_q,
+            cu_seqlens_k=moba_cu_seqlen_kv,
+            max_seqlen_q=max_seqlen,
+            max_seqlen_k=moba_chunk_size,
+            softmax_scale=softmax_scale,
+            causal=False,
+            dropout_p=0.0,
+        )
+
+        self_attn_lse_sh = self_attn_lse_hs.t().contiguous()
+        moba_attn_lse = moba_attn_lse_hs.t().contiguous()
+
+        output = torch.zeros((q.shape[0], q.shape[1], q.shape[2]), device=q.device, dtype=torch.float32)
+        output_2d = output.view(-1, q.shape[2])
+
+        max_lse_1d = self_attn_lse_sh.view(-1)
+        max_lse_1d = max_lse_1d.index_reduce(
+            0, moba_q_sh_indices, moba_attn_lse.view(-1), "amax"
+        )
+        self_attn_lse_sh = self_attn_lse_sh - max_lse_1d.view_as(self_attn_lse_sh)
+        moba_attn_lse = (
+            moba_attn_lse.view(-1)
+            .sub(max_lse_1d.index_select(0, moba_q_sh_indices))
+            .reshape_as(moba_attn_lse)
+        )
+
+        mixed_attn_se_sh = self_attn_lse_sh.exp()
+        moba_attn_se = moba_attn_lse.exp()
+
+        mixed_attn_se_sh.view(-1).index_add_(
+            0, moba_q_sh_indices, moba_attn_se.view(-1)
+        )
+        mixed_attn_lse_sh = mixed_attn_se_sh.log()
+
+        # Combine self-attention output
+        factor = (self_attn_lse_sh - mixed_attn_lse_sh).exp()  # [S, H]
+        self_attn_out_sh = self_attn_out_sh * factor.unsqueeze(-1)
+        output_2d += self_attn_out_sh.reshape_as(output_2d)
+
+        # Combine MOBA attention output
+        mixed_attn_lse = (
+            mixed_attn_lse_sh.view(-1)
+            .index_select(0, moba_q_sh_indices)
+            .view_as(moba_attn_lse)
+        )
+        factor = (moba_attn_lse - mixed_attn_lse).exp()  # [S, H]
+        moba_attn_out = moba_attn_out * factor.unsqueeze(-1)
+        raw_attn_out = moba_attn_out.view(-1, moba_attn_out.shape[-1])
+        output_2d.index_add_(0, moba_q_sh_indices, raw_attn_out)
+        output = output.to(q.dtype)
+        mixed_attn_lse_sh = mixed_attn_lse_sh + max_lse_1d.view_as(mixed_attn_se_sh)
+        ctx.save_for_backward(
+            output,
+            mixed_attn_lse_sh,
+            q,
+            k,
+            v,
+            self_attn_cu_seqlen,
+            moba_q,
+            moba_kv,
+            moba_cu_seqlen_q,
+            moba_cu_seqlen_kv,
+            moba_q_sh_indices,
+        )
+
+        return output
+
+    @staticmethod
+    def backward(ctx, d_output):
+
+        max_seqlen = ctx.max_seqlen
+        moba_chunk_size = ctx.moba_chunk_size
+        softmax_scale = ctx.softmax_scale
+
+        (
+            output,
+            mixed_attn_vlse_sh,
+            q,
+            k,
+            v,
+            self_attn_cu_seqlen,
+            moba_q,
+            moba_kv,
+            moba_cu_seqlen_q,
+            moba_cu_seqlen_kv,
+            moba_q_sh_indices,
+        ) = ctx.saved_tensors
+
+        d_output = d_output.contiguous()
+
+        dq = torch.empty_like(q)
+        dk = torch.empty_like(k)
+        dv = torch.empty_like(v)
+        _ = _flash_attn_varlen_backward(
+            dout=d_output,
+            q=q,
+            k=k,
+            v=v,
+            out=output,
+            softmax_lse=mixed_attn_vlse_sh.t().contiguous(),
+            dq=dq,
+            dk=dk,
+            dv=dv,
+            cu_seqlens_q=self_attn_cu_seqlen,
+            cu_seqlens_k=self_attn_cu_seqlen,
+            max_seqlen_q=max_seqlen,
+            max_seqlen_k=max_seqlen,
+            softmax_scale=softmax_scale,
+            causal=False,
+            dropout_p=0.0,
+            softcap=0.0,
+            alibi_slopes=None,
+            deterministic=True,
+            window_size_left=-1,
+            window_size_right=-1
+        )
+
+        headdim = q.shape[-1]
+        d_moba_output = (
+            d_output.view(-1, headdim).index_select(0, moba_q_sh_indices).unsqueeze(1)
+        )
+        moba_output = (
+            output.view(-1, headdim).index_select(0, moba_q_sh_indices).unsqueeze(1)
+        )
+
+        mixed_attn_vlse = (
+            mixed_attn_vlse_sh.view(-1).index_select(0, moba_q_sh_indices).view(1, -1)
+        )
+
+        dmq = torch.empty_like(moba_q)
+        dmkv = torch.empty_like(moba_kv)
+        _ = _flash_attn_varlen_backward(
+            dout=d_moba_output,
+            q=moba_q,
+            k=moba_kv[:, 0],
+            v=moba_kv[:, 1],
+            out=moba_output,
+            softmax_lse=mixed_attn_vlse,
+            dq=dmq,
+            dk=dmkv[:,0],
+            dv=dmkv[:,1],
+            cu_seqlens_q=moba_cu_seqlen_q,
+            cu_seqlens_k=moba_cu_seqlen_kv,
+            max_seqlen_q=max_seqlen,
+            max_seqlen_k=moba_chunk_size,
+            softmax_scale=softmax_scale,
+            causal=False,
+            dropout_p=0.0,
+            softcap=0.0,
+            alibi_slopes=None,
+            deterministic=True,
+            window_size_left=-1,
+            window_size_right=-1
+        )
+
+        return dq, dk, dv, None, dmq, dmkv, None, None, None, None, None
+
+
+def moba_attn_varlen(
+    q: torch.Tensor,
+    k: torch.Tensor,
+    v: torch.Tensor,
+    cu_seqlens: torch.Tensor,
+    max_seqlen: int,
+    moba_chunk_size: int,
+    moba_topk: int,
+    select_mode: str = 'threshold', # "topk" or "threshold"
+    simsum_threshold: float = 0.25,
+    threshold_type: str = 'query_head',
+) -> torch.Tensor:
+    """
+    Accelerated MOBA attention for vision tasks with proper LSE normalization.
+
+    This version:
+      - Splits KV into chunks.
+      - For each query head, selects the top-k relevant KV chunks (including the self chunk)
+        by amplifying the diagonal (self-chunk) logits.
+      - Aggregates the attention outputs from the selected chunks using a log-sum-exp
+        reduction so that attending to each query over the selected chunks is equivalent
+        to the original algorithm.
+    """
+    # Stack keys and values.
+    kv = torch.stack((k, v), dim=1)
+    seqlen, num_head, head_dim = q.shape
+
+    # Compute chunk boundaries.
+    cu_chunk, filtered_chunk_indices, num_filtered_chunk, chunk_to_batch = calc_chunks(
+        cu_seqlens, moba_chunk_size
+    )
+
+    self_attn_cu_seqlen = cu_chunk
+
+    # Update top-k selection to include the self chunk.
+    moba_topk = min(moba_topk, num_filtered_chunk)
+
+    # --- Build filtered KV from chunks ---
+    chunk_starts = cu_chunk[filtered_chunk_indices]          # [num_filtered_chunk]
+    chunk_ends = cu_chunk[filtered_chunk_indices + 1]          # [num_filtered_chunk]
+    chunk_lengths = chunk_ends - chunk_starts                  # [num_filtered_chunk]
+    max_chunk_len = int(chunk_lengths.max().item())
+
+    range_tensor = torch.arange(max_chunk_len, device=kv.device, dtype=chunk_starts.dtype).unsqueeze(0)
+    indices = chunk_starts.unsqueeze(1) + range_tensor
+    indices = torch.clamp(indices, max=kv.shape[0] - 1)
+    valid_mask = range_tensor < chunk_lengths.unsqueeze(1)
+    gathered = kv[indices.view(-1)].view(num_filtered_chunk, max_chunk_len, *kv.shape[1:])
+    gathered = gathered * valid_mask.unsqueeze(-1).unsqueeze(-1).unsqueeze(-1).type_as(gathered)
+
+    # Compute key_gate_weight over valid tokens.
+    key_values = gathered[:, :, 0].float()  # [num_filtered_chunk, max_chunk_len, num_head, head_dim]
+    valid_mask_exp = valid_mask.unsqueeze(-1).unsqueeze(-1)
+    key_sum = (key_values * valid_mask_exp).sum(dim=1)
+    divisor = valid_mask.sum(dim=1).unsqueeze(-1).unsqueeze(-1)
+    key_gate_weight = key_sum / divisor  # [num_filtered_chunk, num_head, head_dim]
+
+    # Compute gate logits between key_gate_weight and queries.
+    q_float = q.float()
+    # gate = torch.einsum("nhd,shd->nhs", key_gate_weight, q_float)  # [num_filtered_chunk, num_head, seqlen]
+    gate = torch.bmm(key_gate_weight.permute(1, 0, 2), q_float.permute(1, 0, 2).transpose(1, 2)).permute(1, 0, 2)
+
+    # Amplify the diagonal (self chunk) contributions.
+    gate_seq_idx = torch.arange(seqlen, device=q.device, dtype=torch.int32).unsqueeze(0).expand(num_filtered_chunk, seqlen)
+    chunk_start = cu_chunk[filtered_chunk_indices]  # [num_filtered_chunk]
+    chunk_end = cu_chunk[filtered_chunk_indices + 1]  # [num_filtered_chunk]
+    gate_self_chunk_mask = ((gate_seq_idx >= chunk_start.unsqueeze(1)) &
+                            (gate_seq_idx < chunk_end.unsqueeze(1))).unsqueeze(1).expand(-1, num_head, -1)
+    amplification_factor = 1e9  # Example factor; adjust as needed.
+    origin_gate = gate.clone()
+    gate = gate.clone()
+    if select_mode == "topk":
+        gate[gate_self_chunk_mask] += amplification_factor
+
+    # Exclude positions that are outside the valid batch boundaries.
+    batch_starts = cu_seqlens[chunk_to_batch[filtered_chunk_indices]]
+    batch_ends = cu_seqlens[chunk_to_batch[filtered_chunk_indices] + 1]
+    gate_batch_start_mask = gate_seq_idx < batch_starts.unsqueeze(1)
+    gate_batch_end_mask = gate_seq_idx >= batch_ends.unsqueeze(1)
+    gate_inf_mask = gate_batch_start_mask | gate_batch_end_mask
+    gate.masked_fill_(gate_inf_mask.unsqueeze(1), -float("inf"))
+
+    if select_mode == 'topk':
+        # We amplify self‐chunk in gate already, so self entries will rank highest.
+        valid_gate_mask = gate != -float("inf")
+        if threshold_type == 'query_head':
+            # === per‐<head,seq> top-k across chunks (original behavior) ===
+            # gate: (C, H, S)
+            _, gate_topk_idx = torch.topk(gate, k=moba_topk, dim=0, largest=True, sorted=False)
+            gate_idx_mask = torch.zeros_like(gate, dtype=torch.bool)
+            gate_idx_mask.scatter_(0, gate_topk_idx, True)
+            gate_mask = valid_gate_mask & gate_idx_mask
+        elif threshold_type == 'overall':
+            # === global top-k across all (chunk, head, seq) entries ===
+            C, H, S = gate.shape
+            flat_gate = gate.flatten()
+            flat_mask = valid_gate_mask.flatten()
+            flat_gate_masked = torch.where(flat_mask, flat_gate, -float("inf"))
+            # pick topk global entries
+            vals, idx = torch.topk(flat_gate_masked, k=moba_topk * H * S, largest=True, sorted=False)
+            others_mask_flat = torch.zeros_like(flat_mask, dtype=torch.bool)
+            others_mask_flat[idx] = True
+            gate_mask = (valid_gate_mask.flatten() & others_mask_flat).view(gate.shape)
+        elif threshold_type == 'head_global':
+            # per-head top-k across all chunks and sequence positions
+            C, H, S = gate.shape
+            CS = C * S
+            flat_gate  = gate.permute(1, 0, 2).reshape(H, CS)
+            flat_valid = valid_gate_mask.permute(1, 0, 2).reshape(H, CS)
+            flat_gate_masked = torch.where(flat_valid, flat_gate, torch.full_like(flat_gate, -float('inf')))
+            # pick top-k indices per head
+            _, topk_idx = torch.topk(flat_gate_masked, k=moba_topk * S, dim=1, largest=True, sorted=False)
+            gate_idx_flat = torch.zeros_like(flat_valid, dtype=torch.bool)
+            gate_idx_flat.scatter_(1, topk_idx, True)
+            gate_mask = gate_idx_flat.reshape(H, C, S).permute(1, 0, 2)
+        else:
+            raise ValueError(
+                f"Invalid threshold_type for topk: {threshold_type}. "
+                "Choose 'query_head', 'block', or 'overall'."
+            )
+    elif select_mode == 'threshold':
+        # Delegate to the specific thresholding function
+        valid_gate_mask = gate != -float("inf")  # (num_chunk, num_head, seqlen)
+        if threshold_type == 'query_head':
+            gate_mask = _select_threshold_query_head(gate, valid_gate_mask, gate_self_chunk_mask, simsum_threshold)
+        elif threshold_type == 'block':
+            gate_mask = _select_threshold_block(gate, valid_gate_mask, gate_self_chunk_mask, simsum_threshold)
+        elif threshold_type == 'overall':
+            gate_mask = _select_threshold_overall(gate, valid_gate_mask, gate_self_chunk_mask, simsum_threshold)
+        elif threshold_type == 'head_global':
+            gate_mask = _select_threshold_head_global(gate, valid_gate_mask, gate_self_chunk_mask, simsum_threshold)
+        else:
+            raise ValueError(f"Invalid threshold_type: {threshold_type}. Choose 'query_head', 'block', or 'overall'.")
+    else:
+        raise ValueError(f"Invalid select_mode: {select_mode}. Choose 'topk' or 'threshold'.")
+    
+    # eliminate self_chunk in MoBA branch
+    gate_mask = gate_mask & ~gate_self_chunk_mask
+    # if gate_mask is all false, perform flash_attn instead
+    if gate_mask.sum() == 0:
+        return flash_attn_varlen_func(
+            q, k, v, cu_seqlens, cu_seqlens, max_seqlen, max_seqlen, causal=False
+        )
+
+    # Determine which query positions are selected.
+    # nonzero_indices has shape [N, 3] where each row is [chunk_index, head_index, seq_index].
+    moba_q_indices = gate_mask.reshape(gate_mask.shape[0], -1).nonzero(as_tuple=True)[-1] # [(h s k)]
+    moba_q_sh_indices = (moba_q_indices % seqlen) * num_head + (moba_q_indices // seqlen)
+    moba_q = rearrange(q, "s h d -> (h s) d").index_select(0, moba_q_indices).unsqueeze(1)
+
+    # Build cumulative sequence lengths for the selected queries.
+    moba_seqlen_q = gate_mask.sum(dim=-1).flatten()
+    q_zero_mask = moba_seqlen_q == 0
+    valid_expert_mask = ~q_zero_mask
+    if q_zero_mask.sum() > 0:
+        moba_seqlen_q = moba_seqlen_q[valid_expert_mask]
+    moba_cu_seqlen_q = torch.cat(
+        (
+            torch.tensor([0], device=q.device, dtype=moba_seqlen_q.dtype),
+            moba_seqlen_q.cumsum(dim=0),
+        ),
+        dim=0,
+    ).to(torch.int32)
+
+    # Rearrange gathered KV for the MOBA branch.
+    experts_tensor = rearrange(gathered, "nc cl two h d -> (nc h) cl two d")
+    valid_expert_lengths = chunk_lengths.unsqueeze(1).expand(num_filtered_chunk, num_head).reshape(-1).to(torch.int32)
+    if q_zero_mask.sum() > 0:
+        experts_tensor = experts_tensor[valid_expert_mask]
+        valid_expert_lengths = valid_expert_lengths[valid_expert_mask]
+
+    seq_range = torch.arange(experts_tensor.shape[1], device=experts_tensor.device).unsqueeze(0)
+    mask = seq_range < valid_expert_lengths.unsqueeze(1)
+    moba_kv = experts_tensor[mask]  # Shape: ((nc h cl_valid) two d)
+    moba_kv = moba_kv.unsqueeze(2)   # Shape: ((nc h cl_valid) two 1 d)
+
+    moba_cu_seqlen_kv = torch.cat(
+        [torch.zeros(1, device=experts_tensor.device, dtype=torch.int32),
+         valid_expert_lengths.cumsum(dim=0)],
+        dim=0,
+    ).to(torch.int32)
+
+    assert (
+        moba_cu_seqlen_kv.shape == moba_cu_seqlen_q.shape
+    ), f"Mismatch between moba_cu_seqlen_kv.shape and moba_cu_seqlen_q.shape: {moba_cu_seqlen_kv.shape} vs {moba_cu_seqlen_q.shape}"
+
+    return MixedAttention.apply(
+        q,
+        k,
+        v,
+        self_attn_cu_seqlen,
+        moba_q,
+        moba_kv,
+        moba_cu_seqlen_q,
+        moba_cu_seqlen_kv,
+        max_seqlen,
+        moba_chunk_size,
+        moba_q_sh_indices,
+    )
+
+
+def process_moba_input(
+    x,
+    patch_resolution,
+    chunk_size,
+):
+    """
+    Process inputs for the attention function.
+
+    Args:
+        x (torch.Tensor): Input tensor with shape [batch_size, num_patches, num_heads, head_dim].
+        patch_resolution (tuple): Tuple containing the patch resolution (t, h, w).
+        chunk_size (int): Size of the chunk. (maybe tuple or int, according to chunk type)
+
+    Returns:
+        torch.Tensor: Processed input tensor.
+    """
+    if isinstance(chunk_size, float) or isinstance(chunk_size, int):
+        moba_chunk_size = int(chunk_size * patch_resolution[1] * patch_resolution[2])
+    else:
+        assert isinstance(chunk_size, (Tuple, list)), f"chunk_size should be a tuple, list, or int, now it is: {type(chunk_size)}"
+        if len(chunk_size) == 2:
+            assert patch_resolution[1] % chunk_size[0] == 0 and patch_resolution[2] % chunk_size[1] == 0, f"spatial patch_resolution {patch_resolution[1:]} should be divisible by 2d chunk_size {chunk_size}"
+            nch, ncw = patch_resolution[1] // chunk_size[0], patch_resolution[2] // chunk_size[1]
+            x = rearrange(x, "b (t nch ch ncw cw) n d -> b (nch ncw t ch cw) n d", t=patch_resolution[0], nch=nch, ncw=ncw, ch=chunk_size[0], cw=chunk_size[1])
+            moba_chunk_size = patch_resolution[0] * chunk_size[0] * chunk_size[1]
+        elif len(chunk_size) == 3:
+            assert patch_resolution[0] % chunk_size[0] == 0 and patch_resolution[1] % chunk_size[1] == 0 and patch_resolution[2] % chunk_size[2] == 0, f"patch_resolution {patch_resolution} should be divisible by 3d chunk_size {chunk_size}"
+            nct, nch, ncw = patch_resolution[0] // chunk_size[0], patch_resolution[1] // chunk_size[1], patch_resolution[2] // chunk_size[2]
+            x = rearrange(x, "b (nct ct nch ch ncw cw) n d -> b (nct nch ncw ct ch cw) n d", nct=nct, nch=nch, ncw=ncw, ct=chunk_size[0], ch=chunk_size[1], cw=chunk_size[2])
+            moba_chunk_size = chunk_size[0] * chunk_size[1] * chunk_size[2]
+        else:
+            raise ValueError(f"chunk_size should be a int, or a tuple of length 2 or 3, now it is: {len(chunk_size)}")
+        
+    return x, moba_chunk_size
+
+
+def process_moba_output(
+    x,
+    patch_resolution,
+    chunk_size,
+):
+    if isinstance(chunk_size, float) or isinstance(chunk_size, int):
+        pass
+    elif len(chunk_size) == 2:
+        x = rearrange(x, "b (nch ncw t ch cw) n d -> b (t nch ch ncw cw) n d", nch=patch_resolution[1] // chunk_size[0], ncw=patch_resolution[2] // chunk_size[1], t=patch_resolution[0], ch=chunk_size[0], cw=chunk_size[1])
+    elif len(chunk_size) == 3:
+        x = rearrange(x, "b (nct nch ncw ct ch cw) n d -> b (nct ct nch ch ncw cw) n d", nct=patch_resolution[0] // chunk_size[0], nch=patch_resolution[1] // chunk_size[1], ncw=patch_resolution[2] // chunk_size[2], ct=chunk_size[0], ch=chunk_size[1], cw=chunk_size[2])
+
+    return x
+
+
+# TEST
+def generate_data(batch_size, seqlen, num_head, head_dim, dtype):
+    random.seed(0)
+    torch.manual_seed(0)
+    torch.cuda.manual_seed(0)
+    device = torch.cuda.current_device()
+
+    q = torch.randn((batch_size, seqlen, num_head, head_dim), requires_grad=True).to(dtype=dtype, device='cuda')
+    k = torch.randn((batch_size, seqlen, num_head, head_dim), requires_grad=True).to(dtype=dtype, device='cuda')
+    v = torch.randn((batch_size, seqlen, num_head, head_dim), requires_grad=True).to(dtype=dtype, device='cuda')
+    print(f"q.shape: {q.shape}, k.shape: {k.shape}, v.shape: {v.shape}")
+    cu_seqlens = torch.arange(0, q.shape[0] * q.shape[1] + 1, q.shape[1], dtype=torch.int32, device='cuda')
+    max_seqlen = q.shape[1]
+    q = rearrange(q, "b s ... -> (b s) ...")
+    k = rearrange(k, "b s ... -> (b s) ...")
+    v = rearrange(v, "b s ... -> (b s) ...")
+
+    return q, k, v, cu_seqlens, max_seqlen
+
+
+def test_attn_varlen_moba_speed(batch, head, seqlen, head_dim, moba_chunk_size, moba_topk, dtype=torch.bfloat16, select_mode='threshold', simsum_threshold=0.25, threshold_type='query_head'):
+    """Speed test comparing flash_attn vs moba_attention"""
+    # Get data
+    q, k, v, cu_seqlen, max_seqlen = generate_data(batch, seqlen, head, head_dim, dtype)
+    print(f"batch:{batch} head:{head} seqlen:{seqlen} chunk:{moba_chunk_size} topk:{moba_topk} select_mode: {select_mode} simsum_threshold:{simsum_threshold}")
+    vo_grad = torch.randn_like(q)
+    
+    # Warmup
+    warmup_iters = 3
+    perf_test_iters = 10
+
+    # Warmup
+    for _ in range(warmup_iters):
+        o = flash_attn_varlen_func(q, k, v, cu_seqlen, cu_seqlen, max_seqlen, max_seqlen, causal=False)
+        torch.autograd.backward(o, vo_grad)
+    
+    torch.cuda.synchronize()
+    start_flash = time.perf_counter()
+    for _ in range(perf_test_iters):
+        o = flash_attn_varlen_func(q, k, v, cu_seqlen, cu_seqlen, max_seqlen, max_seqlen, causal=False)
+        torch.autograd.backward(o, vo_grad)
+        
+    torch.cuda.synchronize()
+    time_flash = (time.perf_counter() - start_flash) / perf_test_iters * 1000
+
+    # Warmup
+    for _ in range(warmup_iters):
+        om = moba_attn_varlen(q, k, v, cu_seqlen, max_seqlen, moba_chunk_size=moba_chunk_size, moba_topk=moba_topk, select_mode=select_mode, simsum_threshold=simsum_threshold, threshold_type=threshold_type)
+        torch.autograd.backward(om, vo_grad)
+
+    torch.cuda.synchronize()
+    start_moba = time.perf_counter()
+    for _ in range(perf_test_iters):
+        om = moba_attn_varlen(q, k, v, cu_seqlen, max_seqlen, moba_chunk_size=moba_chunk_size, moba_topk=moba_topk, select_mode=select_mode, simsum_threshold=simsum_threshold, threshold_type=threshold_type)
+        torch.autograd.backward(om, vo_grad)
+    
+    torch.cuda.synchronize()
+    time_moba = (time.perf_counter() - start_moba) / perf_test_iters * 1000
+    
+    print(f"Flash: {time_flash:.2f}ms, MoBA: {time_moba:.2f}ms")
+    print(f"Speedup:  {time_flash / time_moba:.2f}x")
+
+
+
+if __name__ == "__main__":
+    """
+    CUDA_VISIBLE_DEVICES=1 \
+    python -u csrc/attn/vmoba_attn/vmoba/vmoba.py
+    """
+    test_attn_varlen_moba_speed(batch=1, head=12, seqlen=32760, head_dim=128, moba_chunk_size=32760 // 3 // 6 // 4, moba_topk=3, select_mode='threshold', simsum_threshold=0.3, threshold_type='query_head')

examples/dataset/vidprom/README.md

@@ -0,0 +1,9 @@
+# VidProm Dataset
+
+From [Self-Forcing](https://github.com/gdhe17/Self-Forcing) repository.
+
+## Download the dataset
+
+```bash
+./download_dataset.sh
+```

examples/dataset/vidprom/download_dataset.sh

@@ -0,0 +1,3 @@
+#! /bin/bash
+
+huggingface-cli download gdhe17/Self-Forcing vidprom_filtered_extended.txt --local-dir prompts

examples/distill/SFWan2.1-T2V/distill_dmd_t2v_1.3B.sh

@@ -0,0 +1,151 @@
+#!/bin/bash
+#SBATCH --job-name=t2v
+#SBATCH --partition=main
+#SBATCH --nodes=1
+#SBATCH --ntasks=1
+#SBATCH --ntasks-per-node=1
+#SBATCH --gres=gpu:1
+#SBATCH --cpus-per-task=128
+#SBATCH --mem=1440G
+#SBATCH --output=dmd_t2v_output/t2v_%j.out
+#SBATCH --error=dmd_t2v_output/t2v_%j.err
+#SBATCH --exclusive
+
+# Basic Info
+export NCCL_P2P_DISABLE=1
+export TORCH_NCCL_ENABLE_MONITORING=0
+# different cache dir for different processes
+export TRITON_CACHE_DIR=/tmp/triton_cache_${SLURM_PROCID}
+export MASTER_PORT=29501
+export TOKENIZERS_PARALLELISM=false
+export WANDB_API_KEY="50632ebd88ffd970521cec9ab4a1a2d7e85bfc45"
+export WANDB_BASE_URL="https://api.wandb.ai"
+export WANDB_MODE=offline
+export FASTVIDEO_ATTENTION_BACKEND=FLASH_ATTN
+
+# Configs
+NUM_GPUS=4
+
+# Model paths for Self-Forcing DMD distillation:
+GENERATOR_MODEL_PATH="wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers"
+REAL_SCORE_MODEL_PATH="Wan-AI/Wan2.1-T2V-14B-Diffusers"  # Teacher model
+FAKE_SCORE_MODEL_PATH="Wan-AI/Wan2.1-T2V-1.3B-Diffusers"  # Critic model
+
+DATA_DIR="data/mixkit-64_processed/Node_0_GPU_1_File_1/combined_parquet_dataset"
+VALIDATION_DATASET_FILE="data/mixkit-64_processed/validation.json"
+# export CUDA_VISIBLE_DEVICES=4,5
+# IP=[MASTER NODE IP]
+
+# Training arguments
+training_args=(
+  --tracker_project_name SFwan_t2v_distill_self_forcing_dmd  # Updated for self-forcing DMD
+  --output_dir "/mnt/sharefs/users/hao.zhang/SFwan_t2v_finetune"
+  --max_train_steps 4000
+  --train_batch_size 1
+  --train_sp_batch_size 1
+  --gradient_accumulation_steps 1
+  --num_latent_t 21
+  --num_height 480
+  --num_width 832
+  --num_frames 81  # Must be divisible by num_frame_per_block (81 % 3 = 0 ✓)
+  --enable_gradient_checkpointing_type "full"
+  --log_visualization
+  --simulate_generator_forward
+  --num_frame_per_block 3  # Frame generation block size for self-forcing
+  --enable_gradient_masking
+  --gradient_mask_last_n_frames 21
+)
+
+# Parallel arguments
+parallel_args=(
+  --num_gpus $NUM_GPUS # 64
+  --sp_size 1
+  --tp_size 1
+  --hsdp_replicate_dim 1 # 64
+  --hsdp_shard_dim $NUM_GPUS
+)
+
+# Model arguments
+model_args=(
+  --model_path $GENERATOR_MODEL_PATH  # TODO: check if you can remove this in this script
+  --pretrained_model_name_or_path $GENERATOR_MODEL_PATH
+  --generator_model_path $GENERATOR_MODEL_PATH
+  --real_score_model_path $REAL_SCORE_MODEL_PATH
+  --fake_score_model_path $FAKE_SCORE_MODEL_PATH
+)
+
+# Dataset arguments
+dataset_args=(
+  --data_path "$DATA_DIR"
+  --dataloader_num_workers 4
+)
+
+# Validation arguments
+validation_args=(
+  --log_validation
+  --validation_dataset_file "$VALIDATION_DATASET_FILE"
+  --validation_steps 100
+  --validation_sampling_steps "4"
+  --validation_guidance_scale "6.0" # not used for dmd inference
+)
+
+# Optimizer arguments
+optimizer_args=(
+  --learning_rate 1e-5
+  --mixed_precision "bf16"
+  --training_state_checkpointing_steps 500
+  --weight_only_checkpointing_steps 500
+  --weight_decay 0.01
+  --betas '0.0,0.999'
+  --max_grad_norm 1.0
+)
+
+# Miscellaneous arguments
+miscellaneous_args=(
+  --inference_mode False
+  --checkpoints_total_limit 3
+  --training_cfg_rate 0.0
+  --dit_precision "fp32"
+  --flow_shift 5
+  --seed 1000
+  --use_ema True
+  --ema_decay 0.99
+  --ema_start_step 100
+  --init_weights_from_safetensors "/mnt/weka/home/hao.zhang/wl/Self-Forcing/diffusers_ode_init/model.safetensors"
+)
+
+# Self-forcing DMD arguments
+dmd_args=(
+  --dmd_denoising_steps '1000,750,500,250'
+  --min_timestep_ratio 0.02
+  --max_timestep_ratio 0.98
+  --dfake_gen_update_ratio 5
+  --real_score_guidance_scale 3.0
+  --fake_score_learning_rate 8e-6
+  --fake_score_betas '0.0,0.999'
+  --warp_denoising_step
+)
+
+# Self-forcing specific arguments
+self_forcing_args=(
+  --independent_first_frame False  # Whether to treat first frame independently
+  --same_step_across_blocks True  # Whether to use same denoising step across all blocks
+  --last_step_only False  # Whether to only use the last denoising step
+  --context_noise 0  # Amount of noise to add during context caching (0 = no noise)
+  --validate_cache_structure False  # Set to True for debugging KV cache issues
+)
+
+torchrun \
+--nnodes 1 \
+--master_port $MASTER_PORT \
+--nproc_per_node $NUM_GPUS \
+    fastvideo/training/wan_self_forcing_distillation_pipeline.py \
+    "${parallel_args[@]}" \
+    "${model_args[@]}" \
+    "${dataset_args[@]}" \
+    "${training_args[@]}" \
+    "${optimizer_args[@]}" \
+    "${validation_args[@]}" \
+    "${miscellaneous_args[@]}" \
+    "${dmd_args[@]}" \
+    "${self_forcing_args[@]}"

examples/distill/SFWan2.1-T2V/download_dataset.sh

@@ -0,0 +1,3 @@
+#!/bin/bash
+
+python scripts/huggingface/download_hf.py --repo_id "wlsaidhi/crush-smol-merged" --local_dir "data/crush-smol" --repo_type "dataset"

examples/distill/SFWan2.1-T2V/preprocess_data.sh

@@ -0,0 +1,24 @@
+#!/bin/bash
+
+GPU_NUM=1 # 2,4,8
+MODEL_PATH="Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
+MODEL_TYPE="wan"
+DATA_MERGE_PATH="data/crush-smol/merge.txt"
+OUTPUT_DIR="data/crush-smol_processed_t2v/"
+
+torchrun --nproc_per_node=$GPU_NUM \
+    fastvideo/pipelines/preprocess/v1_preprocess.py \
+    --model_path $MODEL_PATH \
+    --data_merge_path $DATA_MERGE_PATH \
+    --preprocess_video_batch_size 8 \
+    --seed 42 \
+    --max_height 480 \
+    --max_width 832 \
+    --num_frames 81 \
+    --dataloader_num_workers 0 \
+    --output_dir=$OUTPUT_DIR \
+    --train_fps 16 \
+    --samples_per_file 8 \
+    --flush_frequency 8 \
+    --video_length_tolerance_range 5 \
+    --preprocess_task "t2v" 

examples/distill/Wan2.2-TI2V-5B-Diffusers/crush_smol/distill_dmd_VSA_t2v_5B.sh

@@ -98,6 +98,7 @@ dmd_args=(
 torchrun \
 --nnodes 1 \
 --nproc_per_node $NUM_GPUS \
+--master_port $MASTER_PORT \
     fastvideo/training/wan_distillation_pipeline.py \
     "${parallel_args[@]}" \
     "${model_args[@]}" \

examples/distill/Wan2.2-TI2V-5B-Diffusers/crush_smol/distill_dmd_VSA_t2v_5B_lora.sh

@@ -0,0 +1,112 @@
+#!/bin/bash
+
+# Basic Info
+export WANDB_MODE="online"
+export NCCL_P2P_DISABLE=1
+export TORCH_NCCL_ENABLE_MONITORING=0
+export MASTER_PORT=29501
+export TOKENIZERS_PARALLELISM=false
+export WANDB_BASE_URL="https://api.wandb.ai"
+export WANDB_MODE=online
+export FASTVIDEO_ATTENTION_BACKEND=VIDEO_SPARSE_ATTN
+# export FASTVIDEO_ATTENTION_BACKEND=TORCH_SDPA
+
+# Configs
+NUM_GPUS=1
+MODEL_PATH="Wan-AI/Wan2.2-TI2V-5B-Diffusers"
+DATA_DIR="data/crush-smol_processed_ti2v/combined_parquet_dataset/"
+VALIDATION_DATASET_FILE="examples/distill/Wan2.2-TI2V-5B-Diffusers/crush_smol/validation.json"
+# export CUDA_VISIBLE_DEVICES=4,5
+# IP=[MASTER NODE IP]
+
+# Training arguments
+training_args=(
+  --tracker_project_name wan_t2v_distill_dmd_VSA
+  --output_dir="checkpoints/wan_t2v_finetune"
+  --max_train_steps=4000
+  --train_batch_size=1
+  --train_sp_batch_size 1
+  --gradient_accumulation_steps=1
+  --num_latent_t 31
+  --num_height 704
+  --num_width 1280
+  --num_frames 121
+  --enable_gradient_checkpointing_type "full"
+  --training_state_checkpointing_steps=500
+  --weight_only_checkpointing_steps=500
+  --lora_rank 32
+  --lora_training True
+)
+
+# Parallel arguments
+parallel_args=(
+  --num_gpus 1
+  --sp_size 1
+  --tp_size 1
+  --hsdp_replicate_dim 1
+  --hsdp_shard_dim 1
+)
+
+# Model arguments
+model_args=(
+  --model_path $MODEL_PATH
+  --pretrained_model_name_or_path $MODEL_PATH
+)
+
+# Dataset arguments
+dataset_args=(
+  --data_path "$DATA_DIR"
+  --dataloader_num_workers 4
+)
+
+# Validation arguments
+validation_args=(
+  --log_validation
+  --validation_dataset_file "$VALIDATION_DATASET_FILE"
+  --validation_steps 200
+  --validation_sampling_steps "3"
+  --validation_guidance_scale "6.0" # not used for dmd inference
+)
+
+# Optimizer arguments
+optimizer_args=(
+  --learning_rate=1e-4
+  --mixed_precision="bf16"
+  --weight_decay 0.01
+  --max_grad_norm 1.0
+)
+
+# Miscellaneous arguments
+miscellaneous_args=(
+  --inference_mode False
+  --checkpoints_total_limit 3
+  --training_cfg_rate 0.0
+  --dit_precision "fp32"
+  --ema_start_step 0
+  --flow_shift 8
+  --seed 1000
+)
+
+# DMD arguments
+dmd_args=(
+  --dmd_denoising_steps '1000,757,522'
+  --min_timestep_ratio 0.02
+  --max_timestep_ratio 0.98
+  --generator_update_interval 5
+  --real_score_guidance_scale 3.5
+  --VSA_sparsity 0.8
+)
+
+torchrun \
+--nnodes 1 \
+--nproc_per_node $NUM_GPUS \
+--master_port $MASTER_PORT \
+    fastvideo/training/wan_distillation_pipeline.py \
+    "${parallel_args[@]}" \
+    "${model_args[@]}" \
+    "${dataset_args[@]}" \
+    "${training_args[@]}" \
+    "${optimizer_args[@]}" \
+    "${validation_args[@]}" \
+    "${miscellaneous_args[@]}" \
+    "${dmd_args[@]}"

examples/inference/basic/basic.py

@@ -43,4 +43,4 @@ def main():
 
 
 if __name__ == "__main__":
-    main()
+    main()

examples/training/consistency_finetune/ode_init/finetune_ode_init_16k_vidprom.slurm

@@ -1,5 +1,5 @@
 #!/bin/bash
-#SBATCH --job-name=4e6B8_16kFV_no_warp_ode_vidprom
+#SBATCH --job-name=1e5B2_16kFV_warp_ode_vidprom
 #SBATCH --partition=main
 #SBATCH --nodes=1
 #SBATCH --ntasks=1
@@ -7,8 +7,8 @@
 #SBATCH --gres=gpu:8
 #SBATCH --cpus-per-task=128
 #SBATCH --mem=1440G
-#SBATCH --output=ode_vidprom16k_no_warp/ode_vidprom8b16k_4e-6.out
-#SBATCH --error=ode_vidprom16k_no_warp/ode_vidprom8b16k_4e-6.err
+#SBATCH --output=ode_vidprom16k_warp/Dode_vidprom8b16k_1e-5.out
+#SBATCH --error=ode_vidprom16k_warp/Dode_vidprom8b16k_1e-5.err
 #SBATCH --exclusive
 set -e -x
 
@@ -38,19 +38,21 @@ echo "NODE_RANK: $NODE_RANK"
 
 # MODEL_PATH="wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers"
 MODEL_PATH="Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
-DATA_DIR="/mnt/sharefs/users/hao.zhang/klin/preproc/data/test-ode-preprocessing-16k-t2v-1-3b/"
+DATA_DIR="/mnt/sharefs/users/hao.zhang/klin/preproc/data/test-ode-preprocessing-16k-t2v-1-3b-81/"
 VALIDATION_DATASET_FILE="examples/training/consistency_finetune/ode_init/validation.json"
-NUM_GPUS=2
+NUM_GPUS=8
 # export CUDA_VISIBLE_DEVICES=4,5
 # IP=[MASTER NODE IP]
 
 # Training arguments
 training_args=(
   --tracker_project_name "wan_ode_init"
-  --output_dir "vidprom_8b16k_test_no_warp_4e-6"
+  --output_dir "Dwarp_vidprom_8b16k_test_warp_1e-5"
   --override_transformer_cls_name "CausalWanTransformer3DModel"
-  --wandb_run_name "vidprom_8b16k_wan_ode_init_4e-6"
+  --wandb_run_name "Dwarp_vidprom_8b16k_wan_ode_init_1e-5"
   # --resume_from_checkpoint "ode_init_diffusers/"
+  --warp_denoising_step
+  --log_visualization
   --max_train_steps 6001
   --train_batch_size 1
   --train_sp_batch_size 1
@@ -59,7 +61,7 @@ training_args=(
   --num_height 480
   --num_width 832
   --num_frames 77
-  --dmd_denoising_steps "1000,750,500,0"
+  --dmd_denoising_steps "1000,750,500,250"
   --enable_gradient_checkpointing_type "full"
 )
 
@@ -91,11 +93,12 @@ validation_args=(
   --validation_steps 50
   --validation_sampling_steps "50"
   --validation_guidance_scale "6.0"
+  # --init_weights_from_safetensors "/mnt/weka/home/hao.zhang/wl/Self-Forcing/diffusers_ode_init/model.safetensors"
 )
 
 # Optimizer arguments
 optimizer_args=(
-  --learning_rate 4e-6
+  --learning_rate 1e-5
   --mixed_precision "bf16"
   --checkpointing_steps 500
   --weight_decay 1e-4

examples/training/finetune/wan_t2v_1.3B/crush_smol/preprocess_wan_data_t2v_new.sh

@@ -10,6 +10,7 @@ torchrun --nproc_per_node=$GPU_NUM \
     --model_path $MODEL_PATH \
     --mode preprocess \
     --workload_type t2v \
+    --preprocess.video_loader_type torchvision \
     --preprocess.dataset_type merged \
     --preprocess.dataset_path $DATASET_PATH \
     --preprocess.dataset_output_dir $OUTPUT_DIR \

fastvideo/attention/backends/vmoba.py

@@ -0,0 +1,214 @@
+# SPDX-License-Identifier: Apache-2.0
+
+import re
+from dataclasses import dataclass
+
+import torch
+from einops import rearrange
+from flash_attn.bert_padding import pad_input
+
+from csrc.attn.vmoba_attn.vmoba import (moba_attn_varlen, process_moba_input,
+                                        process_moba_output)
+from fastvideo.attention.backends.abstract import (AttentionBackend,
+                                                   AttentionImpl,
+                                                   AttentionMetadata,
+                                                   AttentionMetadataBuilder)
+from fastvideo.logger import init_logger
+
+logger = init_logger(__name__)
+
+
+class VMOBAAttentionBackend(AttentionBackend):
+
+    accept_output_buffer: bool = True
+
+    @staticmethod
+    def get_name() -> str:
+        return "VMOBA_ATTN"
+
+    @staticmethod
+    def get_impl_cls() -> type["VMOBAAttentionImpl"]:
+        return VMOBAAttentionImpl
+
+    @staticmethod
+    def get_metadata_cls() -> type["VideoMobaAttentionMetadata"]:
+        return VideoMobaAttentionMetadata
+
+    @staticmethod
+    def get_builder_cls() -> type["VideoMobaAttentionMetadataBuilder"]:
+        return VideoMobaAttentionMetadataBuilder
+
+
+@dataclass
+class VideoMobaAttentionMetadata(AttentionMetadata):
+    current_timestep: int
+
+    temporal_chunk_size: int
+    temporal_topk: int
+    spatial_chunk_size: tuple[int, int]
+    spatial_topk: int
+    st_chunk_size: tuple[int, int, int]
+    st_topk: int
+
+    moba_select_mode: str
+    moba_threshold: float
+    moba_threshold_type: str
+    patch_resolution: list[int]
+
+    first_full_step: int = 12
+    first_full_layer: int = 0
+    # temporal_layer -> spatial_layer -> st_layer
+    temporal_layer: int = 1
+    spatial_layer: int = 1
+    st_layer: int = 1
+
+
+class VideoMobaAttentionMetadataBuilder(AttentionMetadataBuilder):
+
+    def __init__(self):
+        pass
+
+    def prepare(self):
+        pass
+
+    def build(  # type: ignore
+        self,
+        current_timestep: int,
+        raw_latent_shape: tuple[int, int, int],
+        patch_size: tuple[int, int, int],
+        temporal_chunk_size: int,
+        temporal_topk: int,
+        spatial_chunk_size: tuple[int, int],
+        spatial_topk: int,
+        st_chunk_size: tuple[int, int, int],
+        st_topk: int,
+        moba_select_mode: str = 'threshold',
+        moba_threshold: float = 0.25,
+        moba_threshold_type: str = 'query_head',
+        device: torch.device = None,
+        first_full_layer: int = 0,
+        first_full_step: int = 12,
+        temporal_layer: int = 1,
+        spatial_layer: int = 1,
+        st_layer: int = 1,
+        **kwargs,
+    ) -> VideoMobaAttentionMetadata:
+        if device is None:
+            device = torch.device("cpu")
+        assert raw_latent_shape[0] % patch_size[0] == 0 and raw_latent_shape[
+            1] % patch_size[1] == 0 and raw_latent_shape[2] % patch_size[
+                2] == 0, f"spatial patch_resolution {raw_latent_shape} should be divisible by patch_size {patch_size}"
+        patch_resolution = [
+            t // pt for t, pt in zip(raw_latent_shape, patch_size, strict=False)
+        ]
+
+        return VideoMobaAttentionMetadata(
+            current_timestep=current_timestep,
+            temporal_chunk_size=temporal_chunk_size,
+            temporal_topk=temporal_topk,
+            spatial_chunk_size=spatial_chunk_size,
+            spatial_topk=spatial_topk,
+            st_chunk_size=st_chunk_size,
+            st_topk=st_topk,
+            moba_select_mode=moba_select_mode,
+            moba_threshold=moba_threshold,
+            moba_threshold_type=moba_threshold_type,
+            patch_resolution=patch_resolution,
+            first_full_layer=first_full_layer,
+            first_full_step=first_full_step,
+            temporal_layer=temporal_layer,
+            spatial_layer=spatial_layer,
+            st_layer=st_layer,
+        )
+
+
+class VMOBAAttentionImpl(AttentionImpl):
+
+    def __init__(self,
+                 num_heads,
+                 head_size,
+                 softmax_scale,
+                 causal=False,
+                 num_kv_heads=None,
+                 prefix="",
+                 **extra_impl_args) -> None:
+        self.prefix = prefix
+        self.layer_idx = self._get_layer_idx(prefix)
+
+    def _get_layer_idx(self, prefix: str) -> int | None:
+        match = re.search(r"blocks\.(\d+)", prefix)
+        if not match:
+            raise ValueError(f"Invalid prefix: {prefix}")
+        return int(match.group(1))
+
+    def forward(
+        self,
+        query: torch.Tensor,
+        key: torch.Tensor,
+        value: torch.Tensor,
+        attn_metadata: AttentionMetadata,
+    ) -> torch.Tensor:
+        """
+        query: [B, L, H, D]
+        key:   [B, L, H, D]
+        value: [B, L, H, D]
+        attn_metadata: AttentionMetadata
+        """
+        batch_size, sequence_length, num_heads, head_dim = query.shape
+
+        # select chunk type according to layer idx:
+        loop_layer_num = attn_metadata.temporal_layer + attn_metadata.spatial_layer + attn_metadata.st_layer
+        moba_layer = self.layer_idx - attn_metadata.first_full_layer
+        if moba_layer % loop_layer_num < attn_metadata.temporal_layer:
+            moba_chunk_size = attn_metadata.temporal_chunk_size
+            moba_topk = attn_metadata.temporal_topk
+        elif moba_layer % loop_layer_num < attn_metadata.temporal_layer + attn_metadata.spatial_layer:
+            moba_chunk_size = attn_metadata.spatial_chunk_size
+            moba_topk = attn_metadata.spatial_topk
+        elif moba_layer % loop_layer_num < attn_metadata.temporal_layer + attn_metadata.spatial_layer + attn_metadata.st_layer:
+            moba_chunk_size = attn_metadata.st_chunk_size
+            moba_topk = attn_metadata.st_topk
+
+        # torch.distributed.breakpoint()
+        query, chunk_size = process_moba_input(query,
+                                               attn_metadata.patch_resolution,
+                                               moba_chunk_size)
+        key, chunk_size = process_moba_input(key,
+                                             attn_metadata.patch_resolution,
+                                             moba_chunk_size)
+        value, chunk_size = process_moba_input(value,
+                                               attn_metadata.patch_resolution,
+                                               moba_chunk_size)
+        max_seqlen = query.shape[1]
+        indices_q = torch.arange(0,
+                                 query.shape[0] * query.shape[1],
+                                 device=query.device)
+        cu_seqlens = torch.arange(0,
+                                  query.shape[0] * query.shape[1] + 1,
+                                  query.shape[1],
+                                  dtype=torch.int32,
+                                  device=query.device)
+        query = rearrange(query, "b s ... -> (b s) ...")
+        key = rearrange(key, "b s ... -> (b s) ...")
+        value = rearrange(value, "b s ... -> (b s) ...")
+
+        # current_timestep=attn_metadata.current_timestep
+        hidden_states = moba_attn_varlen(
+            query,
+            key,
+            value,
+            cu_seqlens=cu_seqlens,
+            max_seqlen=max_seqlen,
+            moba_chunk_size=chunk_size,
+            moba_topk=moba_topk,
+            select_mode=attn_metadata.moba_select_mode,
+            simsum_threshold=attn_metadata.moba_threshold,
+            threshold_type=attn_metadata.moba_threshold_type,
+        )
+        hidden_states = pad_input(hidden_states, indices_q, batch_size,
+                                  sequence_length)
+        hidden_states = process_moba_output(hidden_states,
+                                            attn_metadata.patch_resolution,
+                                            moba_chunk_size)
+
+        return hidden_states

fastvideo/configs/backend/vmoba/wan_1.3B_77_448_832.json

@@ -0,0 +1,16 @@
+{
+    "temporal_chunk_size": 2,
+    "temporal_topk": 2,
+    "spatial_chunk_size": [4, 13],
+    "spatial_topk": 6,
+    "st_chunk_size": [4, 4, 13],
+    "st_topk": 18,
+    "moba_select_mode": "topk",
+    "moba_threshold": 0.25,
+    "moba_threshold_type": "query_head",
+    "first_full_layer": 0,
+    "first_full_step": 12,
+    "temporal_layer": 1,
+    "spatial_layer": 1,
+    "st_layer": 1
+}

fastvideo/configs/backend/vmoba/wan_1.3B_77_480_832.json

@@ -0,0 +1,16 @@
+{
+    "temporal_chunk_size": 2,
+    "temporal_topk": 3,
+    "spatial_chunk_size": [3, 4],
+    "spatial_topk": 20,
+    "st_chunk_size": [4, 6, 4],
+    "st_topk": 15,
+    "moba_select_mode": "threshold",
+    "moba_threshold": 0.25,
+    "moba_threshold_type": "query_head",
+    "first_full_layer": 0,
+    "first_full_step": 12,
+    "temporal_layer": 1,
+    "spatial_layer": 1,
+    "st_layer": 1
+}

fastvideo/configs/configs.py

@@ -32,6 +32,29 @@ class DatasetType(str, Enum):
         return [dataset_type.value for dataset_type in cls]
 
 
+class VideoLoaderType(str, Enum):
+    """
+    Enumeration for different video loaders.
+    """
+    TORCHCODEC = "torchcodec"
+    TORCHVISION = "torchvision"
+
+    @classmethod
+    def from_string(cls, value: str) -> "VideoLoaderType":
+        """Convert string to VideoLoader enum."""
+        try:
+            return cls(value.lower())
+        except ValueError:
+            raise ValueError(
+                f"Invalid video loader: {value}. Must be one of: {', '.join([m.value for m in cls])}"
+            ) from None
+
+    @classmethod
+    def choices(cls) -> list[str]:
+        """Get all available choices as strings for argparse."""
+        return [video_loader.value for video_loader in cls]
+
+
 @dataclasses.dataclass
 class PreprocessConfig:
     """Configuration for preprocessing operations."""
@@ -51,6 +74,7 @@ class PreprocessConfig:
     flush_frequency: int = 256
 
     # Video processing parameters
+    video_loader_type: VideoLoaderType = VideoLoaderType.TORCHCODEC
     max_height: int = 480
     max_width: int = 848
     num_frames: int = 163
@@ -120,6 +144,12 @@ class PreprocessConfig:
                                      help="How often to save to parquet files")
 
         # Video processing parameters
+        preprocess_args.add_argument(
+            f"--{prefix_with_dot}video-loader-type",
+            type=str,
+            choices=VideoLoaderType.choices(),
+            default=PreprocessConfig.video_loader_type.value,
+            help="Type of the video loader")
         preprocess_args.add_argument(f"--{prefix_with_dot}max-height",
                                      type=int,
                                      default=PreprocessConfig.max_height,
@@ -174,6 +204,10 @@ class PreprocessConfig:
         if 'dataset_type' in kwargs and isinstance(kwargs['dataset_type'], str):
             kwargs['dataset_type'] = DatasetType.from_string(
                 kwargs['dataset_type'])
+        if 'video_loader_type' in kwargs and isinstance(
+                kwargs['video_loader_type'], str):
+            kwargs['video_loader_type'] = VideoLoaderType.from_string(
+                kwargs['video_loader_type'])
 
         preprocess_config = cls()
         if not update_config_from_args(

fastvideo/configs/models/dits/base.py

@@ -15,9 +15,13 @@ class DiTArchConfig(ArchConfig):
     reverse_param_names_mapping: dict = field(default_factory=dict)
     lora_param_names_mapping: dict = field(default_factory=dict)
     _supported_attention_backends: tuple[AttentionBackendEnum, ...] = (
-        AttentionBackendEnum.SLIDING_TILE_ATTN, AttentionBackendEnum.SAGE_ATTN,
-        AttentionBackendEnum.FLASH_ATTN, AttentionBackendEnum.TORCH_SDPA,
-        AttentionBackendEnum.VIDEO_SPARSE_ATTN)
+        AttentionBackendEnum.SLIDING_TILE_ATTN,
+        AttentionBackendEnum.SAGE_ATTN,
+        AttentionBackendEnum.FLASH_ATTN,
+        AttentionBackendEnum.TORCH_SDPA,
+        AttentionBackendEnum.VIDEO_SPARSE_ATTN,
+        AttentionBackendEnum.VMOBA_ATTN,
+    )
 
     hidden_size: int = 0
     num_attention_heads: int = 0

fastvideo/configs/models/dits/wanvideo.py

@@ -92,6 +92,9 @@ class WanVideoArchConfig(DiTArchConfig):
     pos_embed_seq_len: int | None = None
     exclude_lora_layers: list[str] = field(default_factory=lambda: ["embedder"])
 
+    # Wan MoE
+    boundary_ratio: float | None = None
+
     # Causal Wan
     local_attn_size: int = -1  # Window size for temporal local attention (-1 indicates global attention)
     sink_size: int = 0  # Size of the attention sink, we keep the first `sink_size` frames unchanged when rolling the KV cache

fastvideo/configs/pipelines/wan.py

@@ -12,13 +12,13 @@ from fastvideo.configs.models.vaes import WanVAEConfig
 from fastvideo.configs.pipelines.base import PipelineConfig
 
 
-def t5_postprocess_text(outputs: BaseEncoderOutput) -> torch.tensor:
-    mask: torch.tensor = outputs.attention_mask
-    hidden_state: torch.tensor = outputs.last_hidden_state
+def t5_postprocess_text(outputs: BaseEncoderOutput) -> torch.Tensor:
+    mask: torch.Tensor = outputs.attention_mask
+    hidden_state: torch.Tensor = outputs.last_hidden_state
     seq_lens = mask.gt(0).sum(dim=1).long()
     assert torch.isnan(hidden_state).sum() == 0
     prompt_embeds = [u[:v] for u, v in zip(hidden_state, seq_lens, strict=True)]
-    prompt_embeds_tensor: torch.tensor = torch.stack([
+    prompt_embeds_tensor: torch.Tensor = torch.stack([
         torch.cat([u, u.new_zeros(512 - u.size(0), u.size(1))])
         for u in prompt_embeds
     ],
@@ -39,12 +39,12 @@ class WanT2V480PConfig(PipelineConfig):
     vae_sp: bool = False
 
     # Denoising stage
-    flow_shift: float | None = 8.0
+    flow_shift: float | None = 3.0
 
     # Text encoding stage
     text_encoder_configs: tuple[EncoderConfig, ...] = field(
         default_factory=lambda: (T5Config(), ))
-    postprocess_text_funcs: tuple[Callable[[BaseEncoderOutput], torch.tensor],
+    postprocess_text_funcs: tuple[Callable[[BaseEncoderOutput], torch.Tensor],
                                   ...] = field(default_factory=lambda:
                                                (t5_postprocess_text, ))
 
@@ -68,7 +68,7 @@ class WanT2V720PConfig(WanT2V480PConfig):
     # WanConfig-specific parameters with defaults
 
     # Denoising stage
-    flow_shift: int = 5
+    flow_shift: float | None = 5.0
 
 
 @dataclass
@@ -82,7 +82,7 @@ class WanI2V480PConfig(WanT2V480PConfig):
         default_factory=CLIPVisionConfig)
     image_encoder_precision: str = "fp32"
 
-    def __post_init__(self):
+    def __post_init__(self) -> None:
         self.vae_config.load_encoder = True
         self.vae_config.load_decoder = True
 
@@ -94,7 +94,7 @@ class WanI2V720PConfig(WanI2V480PConfig):
     # WanConfig-specific parameters with defaults
 
     # Denoising stage
-    flow_shift: int = 5
+    flow_shift: float | None = 5.0
 
 
 @dataclass
@@ -104,40 +104,47 @@ class FastWan2_1_T2V_480P_Config(WanT2V480PConfig):
     # WanConfig-specific parameters with defaults
 
     # Denoising stage
-    flow_shift: int = 8
+    flow_shift: float | None = 8.0
     dmd_denoising_steps: list[int] | None = field(
         default_factory=lambda: [1000, 757, 522])
 
-    def __post_init__(self) -> None:
-        self.vae_config.load_encoder = True
-        self.vae_config.load_decoder = True
-
 
 @dataclass
 class Wan2_2_TI2V_5B_Config(WanT2V480PConfig):
-    flow_shift: int = 5
+    flow_shift: float | None = 5.0
     ti2v_task: bool = True
+    expand_timesteps: bool = True
 
     def __post_init__(self) -> None:
         self.vae_config.load_encoder = True
         self.vae_config.load_decoder = True
+        self.dit_config.expand_timesteps = self.expand_timesteps
 
 
 @dataclass
 class FastWan2_2_TI2V_5B_Config(Wan2_2_TI2V_5B_Config):
-    flow_shift: int = 5
+    flow_shift: float | None = 5.0
     dmd_denoising_steps: list[int] | None = field(
         default_factory=lambda: [1000, 757, 522])
 
 
 @dataclass
 class Wan2_2_T2V_A14B_Config(WanT2V480PConfig):
-    pass
+    flow_shift: float | None = 12.0
+    boundary_ratio: float | None = 0.875
+
+    def __post_init__(self) -> None:
+        self.dit_config.boundary_ratio = self.boundary_ratio
 
 
 @dataclass
-class Wan2_2_I2V_A14B_Config(WanT2V480PConfig):
-    pass
+class Wan2_2_I2V_A14B_Config(WanI2V480PConfig):
+    flow_shift: float | None = 5.0
+    boundary_ratio: float | None = 0.900
+
+    def __post_init__(self) -> None:
+        super().__post_init__()
+        self.dit_config.boundary_ratio = self.boundary_ratio
 
 
 # =============================================
@@ -149,3 +156,4 @@ class SelfForcingWanT2V480PConfig(WanT2V480PConfig):
     flow_shift: float | None = 5.0
     dmd_denoising_steps: list[int] | None = field(
         default_factory=lambda: [1000, 750, 500, 250])
+    warp_denoising_step: bool = True

fastvideo/configs/sample/base.py

@@ -40,6 +40,7 @@ class SamplingParam:
     num_inference_steps: int = 50
     guidance_scale: float = 1.0
     guidance_rescale: float = 0.0
+    boundary_ratio: float | None = None
 
     # TeaCache parameters
     enable_teacache: bool = False
@@ -169,6 +170,12 @@ class SamplingParam:
             default=SamplingParam.guidance_rescale,
             help="Guidance rescale factor",
         )
+        parser.add_argument(
+            "--boundary-ratio",
+            type=float,
+            default=SamplingParam.boundary_ratio,
+            help="Boundary timestep ratio",
+        )
         parser.add_argument(
             "--save-video",
             action="store_true",
@@ -193,6 +200,13 @@ class SamplingParam:
             default=SamplingParam.image_path,
             help="Path to input image for image-to-video generation",
         )
+        parser.add_argument(
+            "--moba-config-path",
+            type=str,
+            default=None,
+            help=
+            "Path to a JSON file containing V-MoBA specific configurations.",
+        )
         parser.add_argument(
             "--return-trajectory-latents",
             action="store_true",

fastvideo/configs/sample/wan.py

@@ -144,18 +144,22 @@ class Wan2_2_TI2V_5B_SamplingParam(Wan2_2_Base_SamplingParam):
 
 @dataclass
 class Wan2_2_T2V_A14B_SamplingParam(Wan2_2_Base_SamplingParam):
-    guidance_scale: float = 4.0
-    guidance_scale_2: float = 3.0
+    guidance_scale: float = 4.0  # high_noise
+    guidance_scale_2: float = 3.0  # low_noise
     num_inference_steps: int = 40
     fps: int = 16
+    # NOTE(will): default boundary timestep is tracked by PipelineConfig, but
+    # can be overridden during sampling
 
 
 @dataclass
 class Wan2_2_I2V_A14B_SamplingParam(Wan2_2_Base_SamplingParam):
-    guidance_scale: float = 3.5
-    guidance_scale_2: float = 3.5
+    guidance_scale: float = 3.5  # high_noise
+    guidance_scale_2: float = 3.5  # low_noise
     num_inference_steps: int = 40
     fps: int = 16
+    # NOTE(will): default boundary timestep is tracked by PipelineConfig, but
+    # can be overridden during sampling
 
 
 # =============================================

fastvideo/dataset/__init__.py

@@ -4,7 +4,7 @@ from torchvision.transforms import Lambda
 
 from fastvideo.dataset.parquet_dataset_map_style import (
     build_parquet_map_style_dataloader)
-from fastvideo.dataset.preprocessing_datasets import VideoCaptionMergedDataset
+from fastvideo.dataset.preprocessing_datasets import VideoCaptionMergedDataset, TextDataset
 from fastvideo.dataset.transform import (CenterCropResizeVideo, Normalize255,
                                          TemporalRandomCrop)
 from fastvideo.dataset.validation_dataset import ValidationDataset
@@ -39,7 +39,13 @@ def getdataset(args) -> VideoCaptionMergedDataset:
                                      seed=args.seed)
 
 
+def gettextdataset(args) -> TextDataset:
+    return TextDataset(data_merge_path=args.data_merge_path,
+                       args=args,
+                       seed=args.seed)
+
+
 __all__ = [
     "build_parquet_map_style_dataloader", "ValidationDataset",
-    "VideoCaptionMergedDataset"
+    "VideoCaptionMergedDataset", "TextDataset"
 ]

fastvideo/dataset/dataloader/parquet_io.py

@@ -0,0 +1,264 @@
+"""
+Utilities for converting preprocessing records (dicts) into Arrow tables and
+writing Parquet datasets in fixed-size chunks.
+
+This module centralizes table construction and Parquet file writing so
+pipelines only need to define their PyArrow schema and produce per-sample
+record dictionaries.
+
+Key APIs:
+- records_to_table(records, schema): Safely convert a list of dictionaries into
+  a pa.Table, casting to the provided schema.
+- ParquetDatasetWriter: Buffer tables and flush to a directory as multiple
+  Parquet files with a fixed number of rows per file. Uses temporary files and
+  atomic rename to avoid partially written outputs.
+"""
+
+from __future__ import annotations
+
+import multiprocessing
+import os
+from concurrent.futures import ProcessPoolExecutor
+from typing import Any
+
+import pyarrow as pa
+import pyarrow.parquet as pq
+
+
+def records_to_table(records: list[dict[str, Any]], schema: pa.Schema) -> pa.Table:
+    """Build a PyArrow table from Python record dicts using an explicit schema.
+
+    Arrow will cast values to the target schema when possible (e.g., promoting
+    Python ints/floats to pa.int64/pa.float64), eliminating hand-written per-
+    field array construction.
+
+    Args:
+        records: List of dictionaries, each representing one row. Keys must
+            match schema field names.
+        schema: Target PyArrow schema. Controls field names and types.
+
+    Returns:
+        pa.Table: In-memory table matching the provided schema. If ``records``
+        is empty, returns an empty table with the given schema.
+    """
+    if not records:
+        return pa.table({}, schema=schema)
+    return pa.Table.from_pylist(records, schema=schema)
+
+
+class ParquetDatasetWriter:
+    """Accumulate tables and flush them to a Parquet directory in fixed-size chunks.
+
+    Behavior:
+    - Writes files under worker-specific subdirectories for parallelism.
+    - Uses temporary files and atomic rename to avoid partial files being left
+      behind on failure.
+    - Only full chunks of ``samples_per_file`` rows are written on each flush;
+      any remainder rows are re-buffered for the next flush.
+
+    Note:
+    - Instances are not meant to be shared across processes. Create one writer
+      per process if using multiprocessing.
+    """
+
+    def __init__(self, out_dir: str, samples_per_file: int, compression: str = "zstd") -> None:
+        """Initialize the dataset writer.
+
+        Args:
+            out_dir: Output directory where Parquet files will be written.
+            samples_per_file: Fixed number of rows per Parquet file.
+            compression: Compression codec passed to ``pyarrow.parquet.write_table``
+                (e.g., ``"zstd"``, ``"snappy"``, ``"gzip"``).
+        """
+        self.out_dir = out_dir
+        self.samples_per_file = max(int(samples_per_file), 1)
+        self.compression = compression
+        os.makedirs(self.out_dir, exist_ok=True)
+        self._tables: list[pa.Table] = []
+
+    def append_table(self, table: pa.Table) -> None:
+        """Append a non-empty table to the internal buffer.
+
+        Args:
+            table: A ``pa.Table`` to buffer. Empty or ``None`` tables are ignored.
+        """
+        if table is None or len(table) == 0:
+            return
+        self._tables.append(table)
+
+    def _combine(self) -> pa.Table | None:
+        """Combine all buffered tables into a single table, if any.
+
+        Returns:
+            A concatenated table, a single table if only one was buffered, or
+            ``None`` if no tables are buffered.
+        """
+        if not self._tables:
+            return None
+        if len(self._tables) == 1:
+            return self._tables[0]
+        return pa.concat_tables(self._tables, promote_options='none')
+
+    def flush(self, num_workers: int | None = None, write_remainder: bool = False) -> int:
+        """Write accumulated tables to disk and clear the written portion.
+
+        Only complete chunks of size ``samples_per_file`` are written. Any
+        remainder rows are kept buffered for the next flush.
+
+        Args:
+            num_workers: Optional override for the number of parallel workers
+                used to write chunks. Defaults to ``min(cpu_count, chunks)``.
+            write_remainder: If True, also write any leftover rows (< samples_per_file)
+                as a final small Parquet file (useful for the last flush at the
+                end of preprocessing).
+
+        Returns:
+            int: Number of rows successfully written in this flush call.
+        """
+        combined = self._combine()
+        self._tables = []
+        if combined is None or len(combined) == 0:
+            return 0
+
+        num_samples = len(combined)
+        total_chunks = num_samples // self.samples_per_file
+        if total_chunks == 0:
+            if not write_remainder:
+                # Not enough to form a full chunk; keep buffered for next round
+                # Re-buffer and return 0 written
+                self._tables = [combined]
+                return 0
+            # Last flush: write the small remainder as a final file in worker_0
+            worker_dir = os.path.join(self.out_dir, "worker_0")
+            os.makedirs(worker_dir, exist_ok=True)
+            # Determine next index
+            num_parquets = 0
+            for _, _, files in os.walk(worker_dir):
+                for file in files:
+                    if file.endswith('.parquet'):
+                        num_parquets += 1
+            chunk_path = os.path.join(worker_dir, f"data_chunk_{num_parquets}.parquet")
+            temp_path = chunk_path + '.tmp'
+            pq.write_table(combined, temp_path, compression=self.compression)
+            if os.path.exists(chunk_path):
+                os.remove(chunk_path)
+            os.rename(temp_path, chunk_path)
+            return num_samples
+
+        # Only write full chunks; keep remainder for next flush
+        written_rows = total_chunks * self.samples_per_file
+        remainder = num_samples - written_rows
+
+        table_to_write = combined.slice(0, written_rows)
+        remainder_table = combined.slice(written_rows, remainder) if remainder > 0 else None
+        if remainder_table is not None and len(remainder_table) > 0:
+            if write_remainder:
+                # Write the remainder as a final small file (worker_0)
+                worker_dir = os.path.join(self.out_dir, "worker_0")
+                os.makedirs(worker_dir, exist_ok=True)
+                num_parquets = 0
+                for _, _, files in os.walk(worker_dir):
+                    for file in files:
+                        if file.endswith('.parquet'):
+                            num_parquets += 1
+                remainder_path = os.path.join(worker_dir,
+                                              f"data_chunk_{num_parquets}.parquet")
+                temp_path = remainder_path + '.tmp'
+                pq.write_table(remainder_table,
+                               temp_path,
+                               compression=self.compression)
+                if os.path.exists(remainder_path):
+                    os.remove(remainder_path)
+                os.rename(temp_path, remainder_path)
+            else:
+                self._tables = [remainder_table]
+
+        # Parallel write by chunk ranges
+        if num_workers is None:
+            num_workers = min(multiprocessing.cpu_count(), max(total_chunks, 1))
+        num_workers = max(int(num_workers), 1)
+        chunks_per_worker = (total_chunks + num_workers - 1) // num_workers
+
+        work_ranges: list[tuple[int, int, pa.Table, int, str, int, str]] = []
+        for worker_id in range(num_workers):
+            start_chunk = worker_id * chunks_per_worker
+            end_chunk = min((worker_id + 1) * chunks_per_worker, total_chunks)
+            if start_chunk < end_chunk:
+                work_ranges.append(
+                    (
+                        start_chunk,
+                        end_chunk,
+                        table_to_write,
+                        worker_id,
+                        self.out_dir,
+                        self.samples_per_file,
+                        self.compression,
+                    )
+                )
+
+        written_total = 0
+        if len(work_ranges) == 1:
+            written_total += _process_chunk_range(work_ranges[0])
+            return written_total
+
+        with ProcessPoolExecutor(max_workers=num_workers) as executor:
+            futures = [executor.submit(_process_chunk_range, args) for args in work_ranges]
+            for f in futures:
+                written_total += f.result()
+        return written_total + (len(remainder_table) if write_remainder and remainder_table is not None else 0)
+
+
+def _process_chunk_range(args: Any) -> int:
+    """Worker function to write a contiguous range of chunk files.
+
+    Args:
+        args: Tuple containing
+            - start_chunk (int): inclusive start chunk index
+            - end_chunk (int): exclusive end chunk index
+            - table (pa.Table): concatenated table containing all rows to write
+            - worker_id (int): numeric worker identifier
+            - output_dir (str): base output directory
+            - samples_per_file (int): rows per chunk file
+            - compression (str): compression codec for Parquet
+
+    Returns:
+        int: Total number of rows written by this worker.
+    """
+    start_chunk, end_chunk, table, worker_id, output_dir, samples_per_file, compression = args
+    total_written = 0
+    num_samples = len(table)
+
+    worker_dir = os.path.join(output_dir, f"worker_{worker_id}")
+    os.makedirs(worker_dir, exist_ok=True)
+
+    # Offset to continue numbering if files exist
+    num_parquets = 0
+    for root, _, files in os.walk(worker_dir):
+        for file in files:
+            if file.endswith('.parquet'):
+                num_parquets += 1
+
+    for i in range(start_chunk, end_chunk):
+        start_sample = i * samples_per_file
+        end_sample = min((i + 1) * samples_per_file, num_samples)
+        if end_sample <= start_sample:
+            continue
+        chunk = table.slice(start_sample, end_sample - start_sample)
+
+        chunk_path = os.path.join(worker_dir, f"data_chunk_{i + num_parquets}.parquet")
+        temp_path = chunk_path + '.tmp'
+        try:
+            pq.write_table(chunk, temp_path, compression=compression)
+            if os.path.exists(chunk_path):
+                os.remove(chunk_path)
+            os.rename(temp_path, chunk_path)
+            total_written += len(chunk)
+        except Exception:
+            if os.path.exists(temp_path):
+                os.remove(temp_path)
+            raise
+
+    return total_written
+
+
+

fastvideo/dataset/dataloader/record_schema.py

@@ -1,5 +1,7 @@
 from typing import Any
 
+import numpy as np
+
 from fastvideo.pipelines.pipeline_batch_info import PreprocessBatch
 
 
@@ -120,3 +122,69 @@ def i2v_record_creator(batch: PreprocessBatch) -> list[dict[str, Any]]:
             })
 
     return records
+
+
+def ode_text_only_record_creator(
+        video_name: str, text_embedding: np.ndarray, caption: str,
+        trajectory_latents: np.ndarray,
+        trajectory_timesteps: np.ndarray) -> dict[str, Any]:
+    """Create a text-only ODE trajectory record matching pyarrow_schema_ode_trajectory_text_only.
+
+    Args:
+        video_name: Base name/id for the sample (without extension).
+        text_embedding: Text encoder output array [SeqLen, Dim].
+        caption: Original text prompt.
+        trajectory_latents: Collected trajectory latents array.
+        trajectory_timesteps: Collected timesteps array.
+
+    Returns:
+        dict suitable for records_to_table(…, pyarrow_schema_ode_trajectory_text_only)
+    """
+    assert trajectory_latents is not None, "trajectory_latents is required"
+    assert trajectory_timesteps is not None, "trajectory_timesteps is required"
+
+    record = {
+        "id": f"text_{video_name}",
+        "text_embedding_bytes": text_embedding.tobytes(),
+        "text_embedding_shape": list(text_embedding.shape),
+        "text_embedding_dtype": str(text_embedding.dtype),
+        "file_name": video_name,
+        "caption": caption,
+        "media_type": "text",
+    }
+
+    record.update({
+        "trajectory_latents_bytes": trajectory_latents.tobytes(),
+        "trajectory_latents_shape": list(trajectory_latents.shape),
+        "trajectory_latents_dtype": str(trajectory_latents.dtype),
+    })
+
+    record.update({
+        "trajectory_timesteps_bytes": trajectory_timesteps.tobytes(),
+        "trajectory_timesteps_shape": list(trajectory_timesteps.shape),
+        "trajectory_timesteps_dtype": str(trajectory_timesteps.dtype),
+    })
+
+    return record
+
+
+def text_only_record_creator(text_name: str, text_embedding: np.ndarray,
+                             caption: str) -> dict[str, Any]:
+    """Create a text-only record matching pyarrow_schema_text_only.
+
+    Args:
+        text_name: Base id/name for the text sample.
+        text_embedding: Text encoder output array [SeqLen, Dim].
+        caption: Original text prompt.
+
+    Returns:
+        dict suitable for records_to_table(…, pyarrow_schema_text_only)
+    """
+    record = {
+        "id": f"text_{text_name}",
+        "text_embedding_bytes": text_embedding.tobytes(),
+        "text_embedding_shape": list(text_embedding.shape),
+        "text_embedding_dtype": str(text_embedding.dtype),
+        "caption": caption,
+    }
+    return record

fastvideo/dataset/dataloader/schema.py

@@ -50,6 +50,7 @@ pyarrow_schema_i2v = pa.schema([
     pa.field("fps", pa.float64()),
 ])
 
+
 pyarrow_schema_t2v = pa.schema([
     pa.field("id", pa.string()),
     # --- Image/Video VAE latents ---
@@ -79,15 +80,9 @@ pyarrow_schema_t2v = pa.schema([
     pa.field("fps", pa.float64()),
 ])
 
-pyarrow_schema_ode_trajectory = pa.schema([
+
+pyarrow_schema_ode_trajectory_text_only = pa.schema([
     pa.field("id", pa.string()),
-    # --- Image/Video VAE latents ---
-    # Tensors are stored as raw bytes with shape and dtype info for loading
-    pa.field("vae_latent_bytes", pa.binary()),
-    # e.g., [C, T, H, W] or [C, H, W]
-    pa.field("vae_latent_shape", pa.list_(pa.int64())),
-    # e.g., 'float32'
-    pa.field("vae_latent_dtype", pa.string()),
     # --- Text encoder output tensor ---
     # Tensors are stored as raw bytes with shape and dtype info for loading
     pa.field("text_embedding_bytes", pa.binary()),
@@ -95,10 +90,6 @@ pyarrow_schema_ode_trajectory = pa.schema([
     pa.field("text_embedding_shape", pa.list_(pa.int64())),
     # e.g., 'bfloat16' or 'float32'
     pa.field("text_embedding_dtype", pa.string()),
-    # I2V
-    pa.field("image_condition_latents_bytes", pa.binary()),
-    pa.field("image_condition_latents_shape", pa.list_(pa.int64())),
-    pa.field("image_condition_latents_dtype", pa.string()),
     # --- ODE Trajectory ---
     pa.field("trajectory_latents_bytes", pa.binary()),
     pa.field("trajectory_latents_shape", pa.list_(pa.int64())),
@@ -109,12 +100,19 @@ pyarrow_schema_ode_trajectory = pa.schema([
     # --- Metadata ---
     pa.field("file_name", pa.string()),
     pa.field("caption", pa.string()),
-    pa.field("media_type", pa.string()),  # 'image' or 'video'
-    pa.field("width", pa.int64()),
-    pa.field("height", pa.int64()),
-    # -- Video-specific (can be null/default for images) ---
-    # Number of frames processed (e.g., 1 for image, N for video)
-    pa.field("num_frames", pa.int64()),
-    pa.field("duration_sec", pa.float64()),
-    pa.field("fps", pa.float64()),
+    pa.field("media_type", pa.string()),  # Always 'text' for text-only
+])
+
+
+pyarrow_schema_text_only = pa.schema([
+    pa.field("id", pa.string()),
+    # --- Text encoder output tensor ---
+    # Tensors are stored as raw bytes with shape and dtype info for loading
+    pa.field("text_embedding_bytes", pa.binary()),
+    # e.g., [SeqLen, Dim]
+    pa.field("text_embedding_shape", pa.list_(pa.int64())),
+    # e.g., 'bfloat16' or 'float32'
+    pa.field("text_embedding_dtype", pa.string()),
+    # --- Metadata ---
+    pa.field("caption", pa.string()),
 ])

fastvideo/dataset/lmdb_datasets.py

@@ -0,0 +1,43 @@
+# SPDX-License-Identifier: Apache-2.0
+from fastvideo.dataset.lmdb_utils import get_array_shape_from_lmdb, retrieve_row_from_lmdb
+from torch.utils.data import Dataset
+import numpy as np
+import torch
+import lmdb
+
+# from Self-Forcing: https://github.com/guandeh17/Self-Forcing/blob/main/utils/dataset.py
+class ODERegressionLMDBDataset(Dataset):
+    def __init__(self, data_path: str, max_pair: int = int(1e8)):
+        print(f"data_path: {data_path}")
+        self.env = lmdb.open(data_path, readonly=True,
+                             lock=False, readahead=False, meminit=False)
+
+        self.latents_shape = get_array_shape_from_lmdb(self.env, 'latents')
+        self.max_pair = max_pair
+
+    def __len__(self):
+        return min(self.latents_shape[0], self.max_pair)
+
+    def __getitem__(self, idx):
+        """
+        Outputs:
+            - prompts: List of Strings
+            - latents: Tensor of shape (num_denoising_steps, num_frames, num_channels, height, width). It is ordered from pure noise to clean image.
+        """
+        latents = retrieve_row_from_lmdb(
+            self.env,
+            "latents", np.float16, idx, shape=self.latents_shape[1:]
+        )
+
+        if len(latents.shape) == 4:
+            latents = latents[None, ...]
+
+        prompts = retrieve_row_from_lmdb(
+            self.env,
+            "prompts", str, idx
+        )
+        return {
+            "prompts": prompts,
+            "ode_latent": torch.tensor(latents, dtype=torch.float32)
+        }
+

fastvideo/dataset/lmdb_utils.py

@@ -0,0 +1,75 @@
+# SPDX-License-Identifier: Apache-2.0
+# from Self-Forcing: https://github.com/guandeh17/Self-Forcing/blob/main/utils/lmdb.py
+
+import numpy as np
+
+
+def get_array_shape_from_lmdb(env, array_name):
+    with env.begin() as txn:
+        image_shape = txn.get(f"{array_name}_shape".encode()).decode()
+        image_shape = tuple(map(int, image_shape.split()))
+    return image_shape
+
+
+def store_arrays_to_lmdb(env, arrays_dict, start_index=0):
+    """
+    Store rows of multiple numpy arrays in a single LMDB.
+    Each row is stored separately with a naming convention.
+    """
+    with env.begin(write=True) as txn:
+        for array_name, array in arrays_dict.items():
+            for i, row in enumerate(array):
+                # Convert row to bytes
+                if isinstance(row, str):
+                    row_bytes = row.encode()
+                else:
+                    row_bytes = row.tobytes()
+
+                data_key = f'{array_name}_{start_index + i}_data'.encode()
+
+                txn.put(data_key, row_bytes)
+
+
+def process_data_dict(data_dict, seen_prompts):
+    output_dict = {}
+
+    all_videos = []
+    all_prompts = []
+    for prompt, video in data_dict.items():
+        if prompt in seen_prompts:
+            continue
+        else:
+            seen_prompts.add(prompt)
+
+        video = video.half().numpy()
+        all_videos.append(video)
+        all_prompts.append(prompt)
+
+    if len(all_videos) == 0:
+        return {"latents": np.array([]), "prompts": np.array([])}
+
+    all_videos = np.concatenate(all_videos, axis=0)
+
+    output_dict['latents'] = all_videos
+    output_dict['prompts'] = np.array(all_prompts)
+
+    return output_dict
+
+
+def retrieve_row_from_lmdb(lmdb_env, array_name, dtype, row_index, shape=None):
+    """
+    Retrieve a specific row from a specific array in the LMDB.
+    """
+    data_key = f'{array_name}_{row_index}_data'.encode()
+
+    with lmdb_env.begin() as txn:
+        row_bytes = txn.get(data_key)
+
+    if dtype == str:
+        array = row_bytes.decode()
+    else:
+        array = np.frombuffer(row_bytes, dtype=dtype)
+
+    if shape is not None and len(shape) > 0:
+        array = array.reshape(shape)
+    return array

fastvideo/dataset/preprocessing_datasets.py

@@ -628,3 +628,134 @@ class VideoCaptionMergedDataset(torch.utils.data.IterableDataset,
     def load_state_dict(self, state_dict: dict[str, Any]) -> None:
         """Load state dict from checkpoint."""
         self.processed_batches = state_dict["processed_batches"]
+
+
+class TextDataset(torch.utils.data.IterableDataset,
+                  torch.distributed.checkpoint.stateful.Stateful):
+    """
+    Text-only dataset for processing prompts from a simple text file.
+    
+    Assumes that data_merge_path is a text file with one prompt per line:
+    A cat playing with a ball
+    A dog running in the park
+    A person cooking dinner
+    ...
+    
+    This dataset processes text data through text encoding stages only.
+    """
+
+    def __init__(self,
+                 data_merge_path: str,
+                 args,
+                 start_idx: int = 0,
+                 seed: int = 42):
+        self.data_merge_path = data_merge_path
+        self.start_idx = start_idx
+        self.args = args
+        self.seed = seed
+
+        # Initialize tokenizer
+        tokenizer_path = os.path.join(args.model_path, "tokenizer")
+        tokenizer = AutoTokenizer.from_pretrained(tokenizer_path,
+                                                  cache_dir=args.cache_dir)
+
+        # Initialize text encoding stage
+        self.text_encoding_stage = TextEncodingStage(
+            tokenizer=tokenizer,
+            text_max_length=args.text_max_length,
+            cfg_rate=getattr(args, 'training_cfg_rate', 0.0),
+            seed=self.seed)
+
+        # Process text data
+        self.processed_batches = self._process_text_data()
+
+    def _load_text_data(self) -> list[str]:
+        """Load text prompts from file."""
+        prompts = []
+        with open(self.data_merge_path, 'r', encoding='utf-8') as f:
+            for line in f:
+                line = line.strip()
+                if line:  # Skip empty lines
+                    prompts.append(line)
+        
+        logger.info(f"Loaded {len(prompts)} text prompts from {self.data_merge_path}")
+        return prompts
+
+    def _process_text_data(self) -> list[PreprocessBatch]:
+        """Process the text prompts through text encoding stage."""
+        raw_prompts = self._load_text_data()
+        processed_batches = []
+
+        for idx, prompt in enumerate(raw_prompts):
+            # Create a text-only batch with dummy path
+            batch = PreprocessBatch(
+                path=f"text_prompt_{idx}",
+                cap=[prompt],  # TextEncodingStage expects a list
+                resolution=None,
+                fps=None,
+                duration=None,
+                num_frames=0,
+                sample_frame_index=None,
+                sample_num_frames=0
+            )
+            
+            processed_batches.append(batch)
+
+        logger.info(f"Processed {len(processed_batches)} text batches")
+        return processed_batches
+
+    def __iter__(self):
+        """Iterator for the dataset."""
+        # Set up distributed sampling if needed
+        if torch.distributed.is_available() and torch.distributed.is_initialized():
+            rank = torch.distributed.get_rank()
+            world_size = torch.distributed.get_world_size()
+        else:
+            rank = 0
+            world_size = 1
+
+        # Calculate chunk for this rank
+        total_items = len(self.processed_batches)
+        items_per_rank = math.ceil(total_items / world_size)
+        start_idx = rank * items_per_rank + self.start_idx
+        end_idx = min(start_idx + items_per_rank, total_items)
+
+        # Yield items for this rank
+        for idx in range(start_idx, end_idx):
+            if idx < len(self.processed_batches):
+                yield self._get_item(idx)
+
+    def _get_item(self, idx: int) -> dict:
+        """Get a single processed text item."""
+        batch = self.processed_batches[idx]
+
+        # Apply text encoding stage
+        batch = self.text_encoding_stage.process(batch)
+
+        # Build result dictionary for text-only processing with required schema fields
+        result = {
+            "text": batch.text,
+            "input_ids": batch.input_ids,
+            "cond_mask": batch.cond_mask,
+            "path": batch.path,
+            # Required schema fields for ODE trajectory processing
+            "id": f"text_{idx}",
+            "file_name": batch.path,
+            "caption": batch.text,
+            "media_type": "text",
+            "width": 1,
+            "height": 1,
+            "num_frames": 0,
+            "duration_sec": 0.0,
+            "fps": 0.0,
+        }
+
+        return result
+
+    def state_dict(self) -> dict[str, Any]:
+        """Return state dict for checkpointing."""
+        return {"processed_batches": self.processed_batches}
+
+    def load_state_dict(self, state_dict: dict[str, Any]) -> None:
+        """Load state dict from checkpoint."""
+        self.processed_batches = state_dict["processed_batches"]

fastvideo/fastvideo_args.py

@@ -1,9 +1,9 @@
 # SPDX-License-Identifier: Apache-2.0
 # Inspired by SGLang: https://github.com/sgl-project/sglang/blob/main/python/sglang/srt/server_args.py
 """The arguments of FastVideo Inference."""
-
 import argparse
 import dataclasses
+import json
 from contextlib import contextmanager
 from dataclasses import field
 from enum import Enum
@@ -139,6 +139,10 @@ class FastVideoArgs:
     # VSA parameters
     VSA_sparsity: float = 0.0  # inference/validation sparsity
 
+    # V-MoBA parameters
+    moba_config_path: str | None = None
+    moba_config: dict[str, Any] = field(default_factory=dict)
+
     # Master port for distributed training/inference
     master_port: int | None = None
 
@@ -167,6 +171,16 @@ class FastVideoArgs:
         return not self.inference_mode
 
     def __post_init__(self):
+        if self.moba_config_path:
+            try:
+                with open(self.moba_config_path) as f:
+                    self.moba_config = json.load(f)
+                logger.info("Loaded V-MoBA config from %s",
+                            self.moba_config_path)
+            except (FileNotFoundError, json.JSONDecodeError) as e:
+                logger.error("Failed to load V-MoBA config from %s: %s",
+                             self.moba_config_path, e)
+                raise
         self.check_fastvideo_args()
 
     @staticmethod
@@ -598,6 +612,11 @@ class TrainingArgs(FastVideoArgs):
     pretrained_model_name_or_path: str = ""
     dit_model_name_or_path: str = ""
 
+    # DMD model paths - separate paths for each network
+    generator_model_path: str = ""  # path for generator (student) model
+    real_score_model_path: str = ""  # path for real score (teacher) model
+    fake_score_model_path: str = ""  # path for fake score (critic) model
+
     # diffusion setting
     ema_decay: float = 0.0
     ema_start_step: int = 0
@@ -620,6 +639,7 @@ class TrainingArgs(FastVideoArgs):
     checkpoints_total_limit: int = 0
     checkpointing_steps: int = 0
     resume_from_checkpoint: str = ""  # specify the checkpoint folder to resume from
+    init_weights_from_safetensors: str = ""  # path to safetensors file for initial weight loading
 
     # optimizer & scheduler
     num_train_epochs: int = 0
@@ -651,6 +671,7 @@ class TrainingArgs(FastVideoArgs):
     linear_quadratic_threshold: float = 0.0
     linear_range: float = 0.0
     weight_decay: float = 0.0
+    betas: str = "0.9,0.999"  # betas for optimizer, format: "beta1,beta2"
     use_ema: bool = False
     multi_phased_distill_schedule: str = ""
     pred_decay_weight: float = 0.0
@@ -671,17 +692,30 @@ class TrainingArgs(FastVideoArgs):
 
     # distillation args
     generator_update_interval: int = 5
+    dfake_gen_update_ratio: int = 5  # self-forcing: how often to train generator vs critic
     min_timestep_ratio: float = 0.2
     max_timestep_ratio: float = 0.98
     real_score_guidance_scale: float = 3.5
     fake_score_learning_rate: float = 0.0  # separate learning rate for fake_score_transformer, if 0.0, use learning_rate
     fake_score_lr_scheduler: str = "constant"  # separate lr scheduler for fake_score_transformer, if not set, use lr_scheduler
+    fake_score_betas: str = "0.9,0.999"  # betas for fake score optimizer, format: "beta1,beta2"
     training_state_checkpointing_steps: int = 0  # for resuming training
     weight_only_checkpointing_steps: int = 0  # for inference
     log_visualization: bool = False
     # simulate generator forward to match inference
     simulate_generator_forward: bool = False
     warp_denoising_step: bool = False
+    intermediate_latents_visualization: bool = False
+
+    # Self-forcing specific arguments
+    num_frame_per_block: int = 3
+    independent_first_frame: bool = False
+    enable_gradient_masking: bool = True
+    gradient_mask_last_n_frames: int = 21
+    validate_cache_structure: bool = False  # Debug flag for cache validation
+    same_step_across_blocks: bool = False  # Use same exit timestep for all blocks
+    last_step_only: bool = False  # Only use the last timestep for training
+    context_noise: int = 0  # Context noise level for cache updates
 
     @classmethod
     def from_cli_args(cls, args: argparse.Namespace) -> "TrainingArgs":
@@ -783,6 +817,20 @@ class TrainingArgs(FastVideoArgs):
                             type=str,
                             help="Directory to cache models")
 
+        # DMD model paths - separate paths for each network
+        parser.add_argument(
+            "--generator-model-path",
+            type=str,
+            help="Path to generator (student) model for DMD distillation")
+        parser.add_argument(
+            "--real-score-model-path",
+            type=str,
+            help="Path to real score (teacher) model for DMD distillation")
+        parser.add_argument(
+            "--fake-score-model-path",
+            type=str,
+            help="Path to fake score (critic) model for DMD distillation")
+
         # Diffusion settings
         parser.add_argument("--ema-decay",
                             type=float,
@@ -853,6 +901,10 @@ class TrainingArgs(FastVideoArgs):
         parser.add_argument("--resume-from-checkpoint",
                             type=str,
                             help="Path to checkpoint to resume from")
+        parser.add_argument(
+            "--init-weights-from-safetensors",
+            type=str,
+            help="Path to safetensors file for initial weight loading")
         parser.add_argument("--logging-dir",
                             type=str,
                             help="Directory for logging")
@@ -957,6 +1009,10 @@ class TrainingArgs(FastVideoArgs):
                             help="Linear quadratic threshold")
         parser.add_argument("--linear-range", type=float, help="Linear range")
         parser.add_argument("--weight-decay", type=float, help="Weight decay")
+        parser.add_argument("--betas",
+                            type=str,
+                            default=TrainingArgs.betas,
+                            help="Betas for optimizer (format: 'beta1,beta2')")
         parser.add_argument("--use-ema",
                             action=StoreBoolean,
                             help="Whether to use EMA")
@@ -993,11 +1049,27 @@ class TrainingArgs(FastVideoArgs):
         parser.add_argument("--lora-rank", type=int, help="LoRA rank")
         parser.add_argument("--lora-alpha", type=int, help="LoRA alpha")
 
+        # V-MoBA parameters
+        parser.add_argument(
+            "--moba-config-path",
+            type=str,
+            default=None,
+            help=
+            "Path to a JSON file containing V-MoBA specific configurations.",
+        )
+
         # Distillation arguments
         parser.add_argument("--generator-update-interval",
                             type=int,
                             default=TrainingArgs.generator_update_interval,
                             help="Ratio of student updates to critic updates.")
+        parser.add_argument(
+            "--dfake-gen-update-ratio",
+            type=int,
+            default=TrainingArgs.dfake_gen_update_ratio,
+            help=
+            "Self-forcing: How often to train generator vs critic (train generator every N steps)."
+        )
         parser.add_argument("--min-timestep-ratio",
                             type=float,
                             default=TrainingArgs.min_timestep_ratio,
@@ -1014,6 +1086,11 @@ class TrainingArgs(FastVideoArgs):
                             type=float,
                             default=TrainingArgs.fake_score_learning_rate,
                             help="Learning rate for fake score transformer")
+        parser.add_argument(
+            "--fake-score-betas",
+            type=str,
+            default=TrainingArgs.fake_score_betas,
+            help="Betas for fake score optimizer (format: 'beta1,beta2')")
         parser.add_argument(
             "--fake-score-lr-scheduler",
             type=str,
@@ -1029,7 +1106,45 @@ class TrainingArgs(FastVideoArgs):
         parser.add_argument(
             "--warp-denoising-step",
             action=StoreBoolean,
-            help="Whether to warp denoising step according to the scheduler time shift")
+            help=
+            "Whether to warp denoising step according to the scheduler time shift"
+        )
+
+        # Self-forcing specific arguments
+        parser.add_argument(
+            "--num-frame-per-block",
+            type=int,
+            default=TrainingArgs.num_frame_per_block,
+            help="Number of frames per block for causal generation")
+        parser.add_argument(
+            "--independent-first-frame",
+            action=StoreBoolean,
+            help="Whether the first frame is independent in causal generation")
+        parser.add_argument(
+            "--enable-gradient-masking",
+            action=StoreBoolean,
+            help="Whether to enable frame-level gradient masking")
+        parser.add_argument(
+            "--gradient-mask-last-n-frames",
+            type=int,
+            default=TrainingArgs.gradient_mask_last_n_frames,
+            help="Number of last frames to enable gradients for")
+        parser.add_argument(
+            "--validate-cache-structure",
+            action=StoreBoolean,
+            help="Whether to validate KV cache structure (debug flag)")
+        parser.add_argument(
+            "--same-step-across-blocks",
+            action=StoreBoolean,
+            help="Whether to use the same exit timestep for all blocks")
+        parser.add_argument(
+            "--last-step-only",
+            action=StoreBoolean,
+            help="Whether to only use the last timestep for training")
+        parser.add_argument("--context-noise",
+                            type=int,
+                            default=TrainingArgs.context_noise,
+                            help="Context noise level for cache updates")
 
         return parser
 

fastvideo/layers/layernorm.py

@@ -9,9 +9,6 @@ import torch.nn.functional as F
 from fastvideo.layers.custom_op import CustomOp
 from fastvideo.platforms import current_platform
 
-from fastvideo.logger import init_logger
-
-logger = init_logger(__name__)
 
 @CustomOp.register("rms_norm")
 class RMSNorm(CustomOp):
@@ -103,13 +100,16 @@ class ScaleResidual(nn.Module):
     def forward(self, residual: torch.Tensor, x: torch.Tensor,
                 gate: torch.Tensor) -> torch.Tensor:
         """Apply gated residual connection."""
-        # logger.info("x.shape: %s", x.shape)
-        # if isinstance(gate, torch.Tensor):
-            # logger.info("gate.shape: %s", gate.shape)
-
-        num_frames = gate.shape[1]
-        frame_seqlen = x.shape[1] // num_frames
-        return residual + (x.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) * gate).flatten(1, 2)
+        # x.shape: [batch_size, seq_len, inner_dim]
+        if gate.dim() == 4:
+            # gate.shape: [batch_size, num_frames, 1, inner_dim]
+            num_frames = gate.shape[1]
+            frame_seqlen = x.shape[1] // num_frames
+            return residual + (x.unflatten(
+                dim=1, sizes=(num_frames, frame_seqlen)) * gate).flatten(1, 2)
+        else:
+            # gate.shape: [batch_size, 1, inner_dim]
+            return residual + x * gate
 
 
 # adapted from Diffusers: https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/normalization.py
@@ -168,7 +168,7 @@ class ScaleResidualLayerNormScaleShift(nn.Module):
             raise NotImplementedError(f"Norm type {norm_type} not implemented")
 
     def forward(self, residual: torch.Tensor, x: torch.Tensor,
-                gate: torch.Tensor, shift: torch.Tensor,
+                gate: torch.Tensor | int, shift: torch.Tensor,
                 scale: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
         """
         Apply gated residual connection, followed by layernorm and 
@@ -180,36 +180,41 @@ class ScaleResidualLayerNormScaleShift(nn.Module):
             - residual value (value after residual connection 
               but before normalization)
         """
+        # x.shape: [batch_size, seq_len, inner_dim]
         # Apply residual connection with gating
-        # logger.info("x.shape: %s", x.shape)
-        if isinstance(gate, int): 
+        if isinstance(gate, int):
             # used by cross-attention, should be 1
             assert gate == 1
-            residual_output = residual + x * gate
+            residual_output = residual + x
         elif isinstance(gate, torch.Tensor):
-            # logger.info("gate.shape: %s", gate.shape)
-            if gate.dim() == 3:
-                # used by bidirectional self attention
-                residual_output = residual + x * gate
-            else:
-                assert gate.dim() == 4 
+            if gate.dim() == 4:
+                # gate.shape: [batch_size, num_frames, 1, inner_dim]
                 num_frames = gate.shape[1]
                 frame_seqlen = x.shape[1] // num_frames
-                residual_output = residual + (x.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) * gate).flatten(1, 2)
-                # residual_output = residual + x * gate
+                residual_output = residual + (
+                    x.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) *
+                    gate).flatten(1, 2)
+            else:
+                # used by bidirectional self attention
+                # gate.shape: [batch_size, 1, inner_dim]
+                residual_output = residual + x * gate
         else:
             raise ValueError(f"Gate type {type(gate)} not supported")
-        # logger.info("residual_output.shape: %s", residual_output.shape)
+        # residual_output.shape: [batch_size, seq_len, inner_dim]
 
         # Apply normalization
         normalized = self.norm(residual_output)
         # Apply scale and shift
         if isinstance(scale, torch.Tensor) and scale.dim() == 4:
+            # scale.shape: [batch_size, num_frames, 1, inner_dim]
+            # shift.shape: [batch_size, num_frames, 1, inner_dim]
             num_frames = scale.shape[1]
             frame_seqlen = normalized.shape[1] // num_frames
-            modulated = (normalized.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) * (1.0 + scale) + shift).flatten(1, 2)
+            modulated = (
+                normalized.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) *
+                (1 + scale) + shift).flatten(1, 2)
         else:
-            modulated = normalized * (1.0 + scale) + shift
+            modulated = normalized * (1 + scale) + shift
         return modulated, residual_output
 
 
@@ -251,16 +256,24 @@ class LayerNormScaleShift(nn.Module):
     def forward(self, x: torch.Tensor, shift: torch.Tensor,
                 scale: torch.Tensor) -> torch.Tensor:
         """Apply ln followed by scale and shift in a single fused operation."""
+        # x.shape: [batch_size, seq_len, inner_dim]
         normalized = self.norm(x)
+        if self.compute_dtype == torch.float32:
+            normalized = normalized.float()
+
         if scale.dim() == 4:
+            # scale.shape: [batch_size, num_frames, 1, inner_dim]
             num_frames = scale.shape[1]
             frame_seqlen = normalized.shape[1] // num_frames
-            if self.compute_dtype == torch.float32:
-                return (normalized.float().unflatten(dim=1, sizes=(num_frames, frame_seqlen)) * (1.0 + scale) + shift).flatten(1, 2).to(x.dtype)
-            else:
-                return (normalized.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) * (1.0 + scale) + shift).flatten(1, 2)
+            output = (
+                normalized.unflatten(dim=1, sizes=(num_frames, frame_seqlen)) *
+                (1 + scale) + shift).flatten(1, 2)
         else:
-            if self.compute_dtype == torch.float32:
-                return (normalized.float() * (1.0 + scale) + shift).to(x.dtype)
-            else:
-                return normalized * (1.0 + scale) + shift
+            # scale.shape: [batch_size, 1, inner_dim]
+            # shift.shape: [batch_size, 1, inner_dim]
+            output = normalized * (1 + scale) + shift
+
+        if self.compute_dtype == torch.float32:
+            output = output.to(x.dtype)
+
+        return output

fastvideo/layers/lora/linear.py

@@ -63,7 +63,7 @@ class BaseLayerWithLoRA(nn.Module):
                             device=self.base_layer.weight.device,
                             dtype=self.base_layer.weight.dtype))
             torch.nn.init.kaiming_uniform_(self.lora_A, a=math.sqrt(5))
-            torch.nn.init.kaiming_uniform_(self.lora_B, a=math.sqrt(5))
+            torch.nn.init.zeros_(self.lora_B)
         else:
             self.lora_A = None
             self.lora_B = None
@@ -77,9 +77,11 @@ class BaseLayerWithLoRA(nn.Module):
             lora_A = self.lora_A.to_local()
 
         if not self.merged and not self.disable_lora:
-            delta = x @ (
-                self.slice_lora_b_weights(lora_B.to(x, non_blocking=True))
-                @ self.slice_lora_a_weights(lora_A.to(x, non_blocking=True)))
+            lora_A_sliced = self.slice_lora_a_weights(
+                lora_A.to(x, non_blocking=True))
+            lora_B_sliced = self.slice_lora_b_weights(
+                lora_B.to(x, non_blocking=True))
+            delta = x @ lora_A_sliced.T @ lora_B_sliced.T
             if self.lora_alpha != self.lora_rank:
                 delta = delta * (
                     self.lora_alpha / self.lora_rank  # type: ignore

fastvideo/models/dits/causal_wanvideo.py

@@ -147,6 +147,9 @@ class CausalWanSelfAttention(nn.Module):
                 # Assign new keys/values directly up to current_end
                 local_end_index = kv_cache["local_end_index"].item() + current_end - kv_cache["global_end_index"].item()
                 local_start_index = local_end_index - num_new_tokens
+                # kv_cache["k"] = kv_cache["k"].detach()
+                # kv_cache["v"] = kv_cache["v"].detach()
+                # logger.info("kv_cache['k'] is in comp graph: %s", kv_cache["k"].requires_grad or kv_cache["k"].grad_fn is not None)
                 kv_cache["k"][:, local_start_index:local_end_index] = roped_key
                 kv_cache["v"][:, local_start_index:local_end_index] = v
             x = self.attn(
@@ -176,7 +179,7 @@ class CausalWanTransformerBlock(nn.Module):
         super().__init__()
 
         # 1. Self-attention
-        self.norm1 = FP32LayerNorm(dim, eps, elementwise_affine=False)
+        self.norm1 = nn.LayerNorm(dim, eps, elementwise_affine=False)
         self.to_q = ReplicatedLinear(dim, dim, bias=True)
         self.to_k = ReplicatedLinear(dim, dim, bias=True)
         self.to_v = ReplicatedLinear(dim, dim, bias=True)
@@ -209,8 +212,7 @@ class CausalWanTransformerBlock(nn.Module):
             norm_type="layer",
             eps=eps,
             elementwise_affine=True,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
+            dtype=torch.float32)
 
         # 2. Cross-attention
         # Only T2V for now
@@ -223,8 +225,7 @@ class CausalWanTransformerBlock(nn.Module):
             norm_type="layer",
             eps=eps,
             elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
+            dtype=torch.float32)
 
         # 3. Feed-forward
         self.ffn = MLP(dim, ffn_dim, act_type="gelu_pytorch_tanh")
@@ -244,40 +245,39 @@ class CausalWanTransformerBlock(nn.Module):
         current_start: int = 0,
         cache_start: int | None = None,
     ) -> torch.Tensor:
-        # logger.info("temb.shape: %s", temb.shape)
-        num_frames = temb.shape[1]
-        # logger.info("first hidden_states.shape: %s", hidden_states.shape)
-        # logger.info("num_frames: %s", num_frames)
+        # hidden_states.shape: [batch_size, seq_length, inner_dim]
+        # temb.shape: [batch_size, num_frames, 6, inner_dim]
         if hidden_states.dim() == 4:
             hidden_states = hidden_states.squeeze(1)
-        frame_seqlen = hidden_states.shape[1] // temb.shape[1]
-        # logger.info("frame_seqlen: %s", frame_seqlen)
+        num_frames = temb.shape[1]
+        frame_seqlen = hidden_states.shape[1] // num_frames    
         bs, seq_length, _ = hidden_states.shape
         orig_dtype = hidden_states.dtype
         # assert orig_dtype != torch.float32
-        e = self.scale_shift_table + temb.float()
-        # logger.info("e.shape: %s", e.shape)
+        e = self.scale_shift_table + temb
+        # e.shape: [batch_size, num_frames, 6, inner_dim]
+        assert e.shape == (bs, num_frames, 6, self.hidden_dim)
         shift_msa, scale_msa, gate_msa, c_shift_msa, c_scale_msa, c_gate_msa = e.chunk(
             6, dim=2)
-        assert shift_msa.dtype == torch.float32
+        # *_msa.shape: [batch_size, num_frames, 1, inner_dim]
+        # assert shift_msa.dtype == torch.float32
+
+        # logger.info("temb sum: %s, dtype: %s", temb.float().sum().item(), temb.dtype)
+        # logger.info("scale_msa sum: %s, dtype: %s", scale_msa.float().sum().item(), scale_msa.dtype)
+        # logger.info("shift_msa sum: %s, dtype: %s", shift_msa.float().sum().item(), shift_msa.dtype)
 
         # 1. Self-attention
-        # print(f"hidden_states: {hidden_states.shape}")
-        # print(f"hidden_states: {scale_msa.shape}")
-        # print(f"hidden_states: {shift_msa.shape}")
-        # norm_hidden_states = (self.norm1(hidden_states.float()) *
-        #                       (1 + scale_msa) + shift_msa).to(orig_dtype)
-        norm_hidden_states = (self.norm1(hidden_states.float()).unflatten(dim=1, sizes=(num_frames, frame_seqlen)) *
-                              (1 + scale_msa) + shift_msa).flatten(1, 2).to(orig_dtype)
-        # logger.info("1 norm_hidden_states.shape: %s", norm_hidden_states.shape)
+        norm_hidden_states = (self.norm1(hidden_states).unflatten(dim=1, sizes=(num_frames, frame_seqlen)) *
+                        (1 + scale_msa) + shift_msa).flatten(1, 2)
+        # logger.info("norm_hidden_states sum: %s, shape: %s", norm_hidden_states.float().sum().item(), norm_hidden_states.shape)
         query, _ = self.to_q(norm_hidden_states)
         key, _ = self.to_k(norm_hidden_states)
         value, _ = self.to_v(norm_hidden_states)
 
         if self.norm_q is not None:
-            query = self.norm_q(query)
+            query = self.norm_q.forward_native(query)
         if self.norm_k is not None:
-            key = self.norm_k(key)
+            key = self.norm_k.forward_native(key)
 
         query = query.squeeze(1).unflatten(2, (self.num_attention_heads, -1))
         key = key.squeeze(1).unflatten(2, (self.num_attention_heads, -1))
@@ -291,10 +291,6 @@ class CausalWanTransformerBlock(nn.Module):
         null_shift = null_scale = torch.tensor([0], device=hidden_states.device)
         norm_hidden_states, hidden_states = self.self_attn_residual_norm(
             hidden_states, attn_output, gate_msa, null_shift, null_scale)
-        # logger.info("after self_attn_residual_norm norm_hidden_states.shape: %s", norm_hidden_states.shape)
-        # logger.info("after self_attn_residual_norm hidden_states.shape: %s", hidden_states.shape)
-        norm_hidden_states, hidden_states = norm_hidden_states.to(
-            orig_dtype), hidden_states.to(orig_dtype)
 
         # 2. Cross-attention
         attn_output = self.attn2(norm_hidden_states,
@@ -303,17 +299,10 @@ class CausalWanTransformerBlock(nn.Module):
                                  crossattn_cache=crossattn_cache)
         norm_hidden_states, hidden_states = self.cross_attn_residual_norm(
             hidden_states, attn_output, 1, c_shift_msa, c_scale_msa)
-        # logger.info("after cross_attn_residual_norm norm_hidden_states.shape: %s", norm_hidden_states.shape)
-        # logger.info("after cross_attn_residual_norm hidden_states.shape: %s", hidden_states.shape)
-        norm_hidden_states, hidden_states = norm_hidden_states.to(
-            orig_dtype), hidden_states.to(orig_dtype)
 
         # 3. Feed-forward
         ff_output = self.ffn(norm_hidden_states)
         hidden_states = self.mlp_residual(hidden_states, ff_output, c_gate_msa)
-        # logger.info("after mlp_residual norm_hidden_states.shape: %s", norm_hidden_states.shape)
-        # logger.info("after mlp_residual hidden_states.shape: %s", hidden_states.shape)
-        hidden_states = hidden_states.to(orig_dtype)
 
         return hidden_states
 
@@ -376,12 +365,9 @@ class CausalWanTransformer3DModel(BaseDiT):
                                             norm_type="layer",
                                             eps=config.eps,
                                             elementwise_affine=False,
-                                            dtype=torch.float32,
-                                            compute_dtype=torch.float32)
-        # Debug: Log configuration values
-        proj_out_dim = config.out_channels * math.prod(config.patch_size)
-        
-        self.proj_out = nn.Linear(inner_dim, proj_out_dim)
+                                            dtype=torch.float32)
+        self.proj_out = nn.Linear(
+            inner_dim, config.out_channels * math.prod(config.patch_size))
         self.scale_shift_table = nn.Parameter(
             torch.randn(1, 2, inner_dim) / inner_dim**0.5)
 
@@ -389,7 +375,7 @@ class CausalWanTransformer3DModel(BaseDiT):
 
         # Causal-specific
         self.block_mask = None
-        self.num_frame_per_block = 1
+        self.num_frame_per_block = 3
         self.independent_first_frame = False
 
         self.__post_init__()
@@ -470,7 +456,6 @@ class CausalWanTransformer3DModel(BaseDiT):
         This function will be run for num_frame times.
         Process the latent frames one by one (1560 tokens each)
         """
-        # logger.info("forward inference hidden_states.shape: %s", hidden_states.shape)
 
         orig_dtype = hidden_states.dtype
         if not isinstance(encoder_hidden_states, torch.Tensor):
@@ -502,15 +487,18 @@ class CausalWanTransformer3DModel(BaseDiT):
         )
         freqs_cos = freqs_cos.to(hidden_states.device)
         freqs_sin = freqs_sin.to(hidden_states.device)
-        freqs_cis = (freqs_cos.float(),
-                     freqs_sin.float()) if freqs_cos is not None else None
+        freqs_cis = (freqs_cos,
+                     freqs_sin) if freqs_cos is not None else None
 
         hidden_states = self.patch_embedding(hidden_states)
+        grid_sizes = torch.stack(
+            [torch.tensor(hidden_states[0].shape[1:], dtype=torch.long)])
         hidden_states = hidden_states.flatten(2).transpose(1, 2)
-        # logger.info("forward inference flattened and transposed hidden_states.shape: %s", hidden_states.shape)
+
+        encoder_hidden_states = torch.cat([encoder_hidden_states, encoder_hidden_states.new_zeros(1, self.text_len - encoder_hidden_states.size(1), encoder_hidden_states.size(2))], dim=1)
 
         temb, timestep_proj, encoder_hidden_states, encoder_hidden_states_image = self.condition_embedder(
-            timestep.flatten(), encoder_hidden_states, encoder_hidden_states_image)
+                        timestep.flatten(), encoder_hidden_states, encoder_hidden_states_image)
         timestep_proj = timestep_proj.unflatten(1, (6, self.hidden_size)).unflatten(dim=0, sizes=timestep.shape)
 
         if encoder_hidden_states_image is not None:
@@ -549,22 +537,15 @@ class CausalWanTransformer3DModel(BaseDiT):
                                         **causal_kwargs)
 
         # 5. Output norm, projection & unpatchify
-        # logger.info("===== INFERENCE 5. Output norm, projection & unpatchify")
-        # logger.info("hidden_states.shape: %s", hidden_states.shape)
-        # logger.info("temb.shape: %s", temb.shape)
-        shift, scale = (self.scale_shift_table + temb.unsqueeze(1)).chunk(2,
-                                                                          dim=1)
+        temb = temb.unflatten(dim=0, sizes=timestep.shape).unsqueeze(2)
+        shift, scale = (self.scale_shift_table.unsqueeze(1) + temb).chunk(2,
+                                                                    dim=2)
         hidden_states = self.norm_out(hidden_states, shift, scale)
         hidden_states = self.proj_out(hidden_states)
 
-        hidden_states = hidden_states.reshape(batch_size, post_patch_num_frames,
-                                              post_patch_height,
-                                              post_patch_width, p_t, p_h, p_w,
-                                              -1)
-        hidden_states = hidden_states.permute(0, 7, 1, 4, 2, 5, 3, 6)
-        output = hidden_states.flatten(6, 7).flatten(4, 5).flatten(2, 3)
+        output = self.unpatchify(hidden_states, grid_sizes)
 
-        return output
+        return torch.stack(output)
 
     def _forward_train(self,
                 hidden_states: torch.Tensor,
@@ -575,8 +556,6 @@ class CausalWanTransformer3DModel(BaseDiT):
                 start_frame: int = 0,
                 **kwargs) -> torch.Tensor:
 
-        # logger.info("===== forward train hidden_states.shape: %s", hidden_states.shape)
-        # logger.info("===== forward train timestep.shape: %s", timestep.shape)
         orig_dtype = hidden_states.dtype
         if not isinstance(encoder_hidden_states, torch.Tensor):
             encoder_hidden_states = encoder_hidden_states[0]
@@ -607,8 +586,8 @@ class CausalWanTransformer3DModel(BaseDiT):
         )
         freqs_cos = freqs_cos.to(hidden_states.device)
         freqs_sin = freqs_sin.to(hidden_states.device)
-        freqs_cis = (freqs_cos.float(),
-                     freqs_sin.float()) if freqs_cos is not None else None
+        freqs_cis = (freqs_cos,
+                     freqs_sin) if freqs_cos is not None else None
 
         # Construct blockwise causal attn mask
         if self.block_mask is None:
@@ -621,14 +600,14 @@ class CausalWanTransformer3DModel(BaseDiT):
             )
 
         hidden_states = self.patch_embedding(hidden_states)
+        grid_sizes = torch.stack(
+            [torch.tensor(hidden_states[0].shape[1:], dtype=torch.long)])
         hidden_states = hidden_states.flatten(2).transpose(1, 2)
-        # logger.info("forward train flattened and transposed hidden_states.shape: %s", hidden_states.shape)
+
+        encoder_hidden_states = torch.cat([encoder_hidden_states, encoder_hidden_states.new_zeros(1, self.text_len - encoder_hidden_states.size(1), encoder_hidden_states.size(2))], dim=1)
 
         temb, timestep_proj, encoder_hidden_states, encoder_hidden_states_image = self.condition_embedder(
-            timestep.flatten(), encoder_hidden_states, encoder_hidden_states_image)
-        # logger.info("forward train timestep_proj.shape: %s", timestep_proj.shape)
-        # logger.info("forward train timestep.shape: %s", timestep.shape)
-        # logger.info("forward train temb.shape: %s", temb.shape)
+                        timestep.flatten(), encoder_hidden_states, encoder_hidden_states_image)
         timestep_proj = timestep_proj.unflatten(1, (6, self.hidden_size)).unflatten(dim=0, sizes=timestep.shape)
 
         if encoder_hidden_states_image is not None:
@@ -649,44 +628,21 @@ class CausalWanTransformer3DModel(BaseDiT):
                     timestep_proj, freqs_cis,
                     block_mask=self.block_mask)
         else:
-            for block_index, block in enumerate(self.blocks):
-                # logger.info("===== TRAIN block %d", block_index)
-                # logger.info("hidden_states.shape: %s", hidden_states.shape)
-                # logger.info("encoder_hidden_states.shape: %s", encoder_hidden_states.shape)
-                # logger.info("timestep_proj.shape: %s", timestep_proj.shape)
-                # logger.info("freqs_cis.shape: %s", freqs_cis.shape)
-                # logger.info("block_mask.shape: %s", self.block_mask.shape)
+            for block in self.blocks:
                 hidden_states = block(hidden_states, encoder_hidden_states,
                                         timestep_proj, freqs_cis,
                                         block_mask=self.block_mask)
 
         # 5. Output norm, projection & unpatchify
-        # logger.info("===== TRAIN 5. Output norm, projection & unpatchify")
-        # logger.info("hidden_states.shape: %s", hidden_states.shape)
-        # logger.info("temb.shape: %s", temb.shape)
-        # shift, scale = (self.scale_shift_table + temb.unsqueeze(1)).chunk(2,
         temb = temb.unflatten(dim=0, sizes=timestep.shape).unsqueeze(2)
-        # logger.info("WTFWTF train temb.shape: %s", temb.shape)
-        # logger.info("WTFWTF train self.scale_shift_table.shape: %s", self.scale_shift_table.shape)
         shift, scale = (self.scale_shift_table.unsqueeze(1) + temb).chunk(2,
-                                                                          dim=2)
-        # logger.info("DEBUG scale.shape: %s", scale.shape)
-        # logger.info("DEBUG shift.shape: %s", shift.shape)
+                                                                    dim=2)
         hidden_states = self.norm_out(hidden_states, shift, scale)
         hidden_states = self.proj_out(hidden_states)
-        # logger.info("DEBUG after proj_out hidden_states.shape: %s", hidden_states.shape)
-        # logger.info(f"DEBUG reshape dimensions: batch_size={batch_size}, post_patch_num_frames={post_patch_num_frames}")
-        # logger.info(f"DEBUG reshape dimensions: post_patch_height={post_patch_height}, post_patch_width={post_patch_width}")
-        # logger.info(f"DEBUG patch dimensions: p_t={p_t}, p_h={p_h}, p_w={p_w}")
 
-        hidden_states = hidden_states.reshape(batch_size, post_patch_num_frames,
-                                              post_patch_height,
-                                              post_patch_width, p_t, p_h, p_w,
-                                              -1)
-        hidden_states = hidden_states.permute(0, 7, 1, 4, 2, 5, 3, 6)
-        output = hidden_states.flatten(6, 7).flatten(4, 5).flatten(2, 3)
+        output = self.unpatchify(hidden_states, grid_sizes)
 
-        return output
+        return torch.stack(output)
 
     def forward(
         self,
@@ -697,3 +653,30 @@ class CausalWanTransformer3DModel(BaseDiT):
             return self._forward_inference(*args, **kwargs)
         else:
             return self._forward_train(*args, **kwargs)
+
+
+    def unpatchify(self, x, grid_sizes):
+        r"""
+
+
+        Args:
+            x (List[Tensor]):
+                List of patchified features, each with shape [L, C_out * prod(patch_size)]
+            grid_sizes (Tensor):
+                Original spatial-temporal grid dimensions before patching,
+
+
+        Returns:
+            Tensor:
+                Reconstructed video tensors with shape [B, C_out, F, H / 8, W / 8]
+        """
+
+        c = self.out_channels
+        out = []
+        for u, v in zip(x, grid_sizes.tolist()):
+            u = u[:math.prod(v)].view(*v, *self.patch_size, c)
+            u = u.permute(6, 0, 3, 1, 4, 2, 5)
+            # u = torch.einsum('fhwpqrc->cfphqwr', u.contiguous())
+            u = u.reshape(c, *[i * j for i, j in zip(v, self.patch_size)])
+            out.append(u)
+        return out

fastvideo/models/dits/wanvideo.py

@@ -1,3 +1,5 @@
+import torch
+import torch.nn as nn
 # SPDX-License-Identifier: Apache-2.0
 
 import math
@@ -37,16 +39,14 @@ class WanImageEmbedding(torch.nn.Module):
     def __init__(self, in_features: int, out_features: int):
         super().__init__()
 
-        self.norm1 = FP32LayerNorm(in_features)
+        self.norm1 = nn.LayerNorm(in_features)
         self.ff = MLP(in_features, in_features, out_features, act_type="gelu")
-        self.norm2 = FP32LayerNorm(out_features)
+        self.norm2 = nn.LayerNorm(out_features)
 
-    def forward(self,
-                encoder_hidden_states_image: torch.Tensor) -> torch.Tensor:
-        dtype = encoder_hidden_states_image.dtype
+    def forward(self, encoder_hidden_states_image: torch.Tensor) -> torch.Tensor:
         hidden_states = self.norm1(encoder_hidden_states_image)
         hidden_states = self.ff(hidden_states)
-        hidden_states = self.norm2(hidden_states).to(dtype)
+        hidden_states = self.norm2(hidden_states)
         return hidden_states
 
 
@@ -62,7 +62,7 @@ class WanTimeTextImageEmbedding(nn.Module):
         super().__init__()
 
         self.time_embedder = TimestepEmbedder(
-            dim, frequency_embedding_size=time_freq_dim, act_layer="silu")
+            dim, frequency_embedding_size=time_freq_dim, act_layer="silu", freq_dtype=torch.float64)
         self.time_modulation = ModulateProjection(dim,
                                                   factor=6,
                                                   act_layer="silu")
@@ -156,12 +156,12 @@ class WanT2VCrossAttention(WanSelfAttention):
         b, n, d = x.size(0), self.num_heads, self.head_dim
 
         # compute query, key, value
-        q = self.norm_q(self.to_q(x)[0]).view(b, -1, n, d)
+        q = self.norm_q.forward_native(self.to_q(x)[0]).view(b, -1, n, d)
 
         if crossattn_cache is not None:
             if not crossattn_cache["is_init"]:
                 crossattn_cache["is_init"] = True
-                k = self.norm_k(self.to_k(context)[0]).view(b, -1, n, d)
+                k = self.norm_k.forward_native(self.to_k(context)[0]).view(b, -1, n, d)
                 v = self.to_v(context)[0].view(b, -1, n, d)
                 crossattn_cache["k"] = k
                 crossattn_cache["v"] = v
@@ -169,7 +169,7 @@ class WanT2VCrossAttention(WanSelfAttention):
                 k = crossattn_cache["k"]
                 v = crossattn_cache["v"]
         else:
-            k = self.norm_k(self.to_k(context)[0]).view(b, -1, n, d)
+            k = self.norm_k.forward_native(self.to_k(context)[0]).view(b, -1, n, d)
             v = self.to_v(context)[0].view(b, -1, n, d)
 
         # compute attention
@@ -213,10 +213,10 @@ class WanI2VCrossAttention(WanSelfAttention):
         b, n, d = x.size(0), self.num_heads, self.head_dim
 
         # compute query, key, value
-        q = self.norm_q(self.to_q(x)[0]).view(b, -1, n, d)
-        k = self.norm_k(self.to_k(context)[0]).view(b, -1, n, d)
+        q = self.norm_q.forward_native(self.to_q(x)[0]).view(b, -1, n, d)
+        k = self.norm_k.forward_native(self.to_k(context)[0]).view(b, -1, n, d)
         v = self.to_v(context)[0].view(b, -1, n, d)
-        k_img = self.norm_added_k(self.add_k_proj(context_img)[0]).view(
+        k_img = self.norm_added_k.forward_native(self.add_k_proj(context_img)[0]).view(
             b, -1, n, d)
         v_img = self.add_v_proj(context_img)[0].view(b, -1, n, d)
         img_x = self.attn(q, k_img, v_img)
@@ -247,7 +247,7 @@ class WanTransformerBlock(nn.Module):
         super().__init__()
 
         # 1. Self-attention
-        self.norm1 = FP32LayerNorm(dim, eps, elementwise_affine=False)
+        self.norm1 = nn.LayerNorm(dim, eps, elementwise_affine=False)
         self.to_q = ReplicatedLinear(dim, dim, bias=True)
         self.to_k = ReplicatedLinear(dim, dim, bias=True)
         self.to_v = ReplicatedLinear(dim, dim, bias=True)
@@ -278,29 +278,29 @@ class WanTransformerBlock(nn.Module):
             norm_type="layer",
             eps=eps,
             elementwise_affine=True,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
+            dtype=torch.float32)
 
         # 2. Cross-attention
         if added_kv_proj_dim is not None:
             # I2V
-            self.attn2 = WanI2VCrossAttention(dim,
-                                              num_heads,
-                                              qk_norm=qk_norm,
+            self.attn2 = WanI2VCrossAttention(dim, 
+                                              num_heads, 
+                                              qk_norm=qk_norm, 
                                               eps=eps)
+
         else:
             # T2V
-            self.attn2 = WanT2VCrossAttention(dim,
-                                              num_heads,
-                                              qk_norm=qk_norm,
+            self.attn2 = WanT2VCrossAttention(dim, 
+                                              num_heads, 
+                                              qk_norm=qk_norm, 
                                               eps=eps)
+
         self.cross_attn_residual_norm = ScaleResidualLayerNormScaleShift(
-            dim,
-            norm_type="layer",
-            eps=eps,
-            elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
+            dim, 
+            norm_type="layer", 
+            eps=eps, 
+            elementwise_affine=False, 
+            dtype=torch.float32)
 
         # 3. Feed-forward
         self.ffn = MLP(dim, ffn_dim, act_type="gelu_pytorch_tanh")
@@ -319,12 +319,11 @@ class WanTransformerBlock(nn.Module):
             hidden_states = hidden_states.squeeze(1)
         bs, seq_length, _ = hidden_states.shape
         orig_dtype = hidden_states.dtype
-        # assert orig_dtype != torch.float32
 
         if temb.dim() == 4:
             # temb: batch_size, seq_len, 6, inner_dim (wan2.2 ti2v)
             shift_msa, scale_msa, gate_msa, c_shift_msa, c_scale_msa, c_gate_msa = (
-                self.scale_shift_table.unsqueeze(0) + temb.float()
+                self.scale_shift_table.unsqueeze(0) + temb
             ).chunk(6, dim=2)
             # batch_size, seq_len, 1, inner_dim
             shift_msa = shift_msa.squeeze(2)
@@ -335,22 +334,20 @@ class WanTransformerBlock(nn.Module):
             c_gate_msa = c_gate_msa.squeeze(2)
         else:
             # temb: batch_size, 6, inner_dim (wan2.1/wan2.2 14B)
-            e = self.scale_shift_table + temb.float()
+            e = self.scale_shift_table + temb
             shift_msa, scale_msa, gate_msa, c_shift_msa, c_scale_msa, c_gate_msa = e.chunk(
                 6, dim=1)
-        assert shift_msa.dtype == torch.float32
 
         # 1. Self-attention
-        norm_hidden_states = (self.norm1(hidden_states.float()) *
-                              (1 + scale_msa) + shift_msa).to(orig_dtype)
+        norm_hidden_states = self.norm1(hidden_states) * (1 + scale_msa) + shift_msa
         query, _ = self.to_q(norm_hidden_states)
         key, _ = self.to_k(norm_hidden_states)
         value, _ = self.to_v(norm_hidden_states)
 
         if self.norm_q is not None:
-            query = self.norm_q(query)
+            query = self.norm_q.forward_native(query)
         if self.norm_k is not None:
-            key = self.norm_k(key)
+            key = self.norm_k.forward_native(key)
 
         query = query.squeeze(1).unflatten(2, (self.num_attention_heads, -1))
         key = key.squeeze(1).unflatten(2, (self.num_attention_heads, -1))
@@ -370,26 +367,20 @@ class WanTransformerBlock(nn.Module):
         null_shift = null_scale = torch.tensor([0], device=hidden_states.device)
         norm_hidden_states, hidden_states = self.self_attn_residual_norm(
             hidden_states, attn_output, gate_msa, null_shift, null_scale)
-        norm_hidden_states, hidden_states = norm_hidden_states.to(
-            orig_dtype), hidden_states.to(orig_dtype)
 
         # 2. Cross-attention
-        attn_output = self.attn2(norm_hidden_states,
-                                 context=encoder_hidden_states,
+        attn_output = self.attn2(norm_hidden_states, 
+                                 context=encoder_hidden_states, 
                                  context_lens=None)
         norm_hidden_states, hidden_states = self.cross_attn_residual_norm(
             hidden_states, attn_output, 1, c_shift_msa, c_scale_msa)
-        norm_hidden_states, hidden_states = norm_hidden_states.to(
-            orig_dtype), hidden_states.to(orig_dtype)
 
         # 3. Feed-forward
         ff_output = self.ffn(norm_hidden_states)
         hidden_states = self.mlp_residual(hidden_states, ff_output, c_gate_msa)
-        hidden_states = hidden_states.to(orig_dtype)
 
         return hidden_states
 
-
 class WanTransformerBlock_VSA(nn.Module):
 
     def __init__(self,
@@ -406,7 +397,7 @@ class WanTransformerBlock_VSA(nn.Module):
         super().__init__()
 
         # 1. Self-attention
-        self.norm1 = FP32LayerNorm(dim, eps, elementwise_affine=False)
+        self.norm1 = nn.LayerNorm(dim, eps, elementwise_affine=False)
         self.to_q = ReplicatedLinear(dim, dim, bias=True)
         self.to_k = ReplicatedLinear(dim, dim, bias=True)
         self.to_v = ReplicatedLinear(dim, dim, bias=True)
@@ -438,8 +429,7 @@ class WanTransformerBlock_VSA(nn.Module):
             norm_type="layer",
             eps=eps,
             elementwise_affine=True,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
+            dtype=torch.float32)
 
         # 2. Cross-attention
         if added_kv_proj_dim is not None:
@@ -459,8 +449,7 @@ class WanTransformerBlock_VSA(nn.Module):
             norm_type="layer",
             eps=eps,
             elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
+            dtype=torch.float32)
 
         # 3. Feed-forward
         self.ffn = MLP(dim, ffn_dim, act_type="gelu_pytorch_tanh")
@@ -480,23 +469,22 @@ class WanTransformerBlock_VSA(nn.Module):
         bs, seq_length, _ = hidden_states.shape
         orig_dtype = hidden_states.dtype
         # assert orig_dtype != torch.float32
-        e = self.scale_shift_table + temb.float()
+        e = self.scale_shift_table + temb
         shift_msa, scale_msa, gate_msa, c_shift_msa, c_scale_msa, c_gate_msa = e.chunk(
             6, dim=1)
-        assert shift_msa.dtype == torch.float32
 
         # 1. Self-attention
-        norm_hidden_states = (self.norm1(hidden_states.float()) *
-                              (1 + scale_msa) + shift_msa).to(orig_dtype)
+        norm_hidden_states = (self.norm1(hidden_states) *
+                              (1 + scale_msa) + shift_msa)
         query, _ = self.to_q(norm_hidden_states)
         key, _ = self.to_k(norm_hidden_states)
         value, _ = self.to_v(norm_hidden_states)
         gate_compress, _ = self.to_gate_compress(norm_hidden_states)
 
         if self.norm_q is not None:
-            query = self.norm_q(query)
+            query = self.norm_q.forward_native(query)
         if self.norm_k is not None:
-            key = self.norm_k(key)
+            key = self.norm_k.forward_native(key)
 
         query = query.squeeze(1).unflatten(2, (self.num_attention_heads, -1))
         key = key.squeeze(1).unflatten(2, (self.num_attention_heads, -1))
@@ -521,8 +509,6 @@ class WanTransformerBlock_VSA(nn.Module):
         null_shift = null_scale = torch.tensor([0], device=hidden_states.device)
         norm_hidden_states, hidden_states = self.self_attn_residual_norm(
             hidden_states, attn_output, gate_msa, null_shift, null_scale)
-        norm_hidden_states, hidden_states = norm_hidden_states.to(
-            orig_dtype), hidden_states.to(orig_dtype)
 
         # 2. Cross-attention
         attn_output = self.attn2(norm_hidden_states,
@@ -530,17 +516,15 @@ class WanTransformerBlock_VSA(nn.Module):
                                  context_lens=None)
         norm_hidden_states, hidden_states = self.cross_attn_residual_norm(
             hidden_states, attn_output, 1, c_shift_msa, c_scale_msa)
-        norm_hidden_states, hidden_states = norm_hidden_states.to(
-            orig_dtype), hidden_states.to(orig_dtype)
 
         # 3. Feed-forward
         ff_output = self.ffn(norm_hidden_states)
         hidden_states = self.mlp_residual(hidden_states, ff_output, c_gate_msa)
-        hidden_states = hidden_states.to(orig_dtype)
 
         return hidden_states
 
 
+
 class WanTransformer3DModel(CachableDiT):
     _fsdp_shard_conditions = WanVideoConfig()._fsdp_shard_conditions
     _compile_conditions = WanVideoConfig()._compile_conditions
@@ -598,8 +582,7 @@ class WanTransformer3DModel(CachableDiT):
                                             norm_type="layer",
                                             eps=config.eps,
                                             elementwise_affine=False,
-                                            dtype=torch.float32,
-                                            compute_dtype=torch.float32)
+                                            dtype=torch.float32)
         self.proj_out = nn.Linear(
             inner_dim, config.out_channels * math.prod(config.patch_size))
         self.scale_shift_table = nn.Parameter(
@@ -659,10 +642,12 @@ class WanTransformer3DModel(CachableDiT):
             rope_theta=10000)
         freqs_cos = freqs_cos.to(hidden_states.device)
         freqs_sin = freqs_sin.to(hidden_states.device)
-        freqs_cis = (freqs_cos.float(),
-                     freqs_sin.float()) if freqs_cos is not None else None
+        freqs_cis = (freqs_cos,
+                     freqs_sin) if freqs_cos is not None else None
 
         hidden_states = self.patch_embedding(hidden_states)
+        grid_sizes = torch.stack(
+            [torch.tensor(hidden_states[0].shape[1:], dtype=torch.long)])
         hidden_states = hidden_states.flatten(2).transpose(1, 2)
 
         # timestep shape: batch_size, or batch_size, seq_len (wan 2.2 ti2v)
@@ -672,6 +657,8 @@ class WanTransformer3DModel(CachableDiT):
         else:
             ts_seq_len = None
 
+        encoder_hidden_states = torch.cat([encoder_hidden_states, encoder_hidden_states.new_zeros(1, self.text_len - encoder_hidden_states.size(1), encoder_hidden_states.size(2))], dim=1)
+
         temb, timestep_proj, encoder_hidden_states, encoder_hidden_states_image = self.condition_embedder(
             timestep, encoder_hidden_states, encoder_hidden_states_image, timestep_seq_len=ts_seq_len)
         if ts_seq_len is not None:
@@ -728,14 +715,35 @@ class WanTransformer3DModel(CachableDiT):
         hidden_states = self.norm_out(hidden_states, shift, scale)
         hidden_states = self.proj_out(hidden_states)
 
-        hidden_states = hidden_states.reshape(batch_size, post_patch_num_frames,
-                                              post_patch_height,
-                                              post_patch_width, p_t, p_h, p_w,
-                                              -1)
-        hidden_states = hidden_states.permute(0, 7, 1, 4, 2, 5, 3, 6)
-        output = hidden_states.flatten(6, 7).flatten(4, 5).flatten(2, 3)
+        output = self.unpatchify(hidden_states, grid_sizes)
 
-        return output
+        return torch.stack(output)
+    
+    def unpatchify(self, x, grid_sizes):
+        r"""
+
+
+        Args:
+            x (List[Tensor]):
+                List of patchified features, each with shape [L, C_out * prod(patch_size)]
+            grid_sizes (Tensor):
+                Original spatial-temporal grid dimensions before patching,
+
+
+        Returns:
+            Tensor:
+                Reconstructed video tensors with shape [B, C_out, F, H / 8, W / 8]
+        """
+
+        c = self.out_channels
+        out = []
+        for u, v in zip(x, grid_sizes.tolist()):
+            u = u[:math.prod(v)].view(*v, *self.patch_size, c)
+            u = u.permute(6, 0, 3, 1, 4, 2, 5)
+            # u = torch.einsum('fhwpqrc->cfphqwr', u.contiguous())
+            u = u.reshape(c, *[i * j for i, j in zip(v, self.patch_size)])
+            out.append(u)
+        return out
 
     def maybe_cache_states(self, hidden_states: torch.Tensor,
                            original_hidden_states: torch.Tensor) -> None:
@@ -827,5 +835,4 @@ class WanTransformer3DModel(CachableDiT):
         if self.is_even:
             return hidden_states + self.previous_residual_even
         else:
-            return hidden_states + self.previous_residual_odd
-        
+            return hidden_states + self.previous_residual_odd

fastvideo/models/loader/component_loader.py

@@ -434,6 +434,16 @@ class TransformerLoader(ComponentLoader):
         if not safetensors_list:
             raise ValueError(f"No safetensors files found in {model_path}")
 
+        # Check if we should use custom initialization weights
+        custom_weights_path = getattr(fastvideo_args, 'init_weights_from_safetensors', None)
+        use_custom_weights = (custom_weights_path and os.path.exists(custom_weights_path) and 
+                            fastvideo_args.training_mode and 
+                            not hasattr(fastvideo_args, '_loading_teacher_critic_model'))
+
+        if use_custom_weights:
+            logger.info("Using custom initialization weights from: %s", custom_weights_path)
+            safetensors_list = [custom_weights_path]
+
         logger.info("Loading model from %s safetensors files in %s",
                     len(safetensors_list), model_path)
 

fastvideo/models/registry.py

@@ -61,6 +61,9 @@ _SCHEDULERS = {
      "FlowMatchEulerDiscreteScheduler"),
     "UniPCMultistepScheduler":
     ("schedulers", "scheduling_unipc_multistep", "UniPCMultistepScheduler"),
+    "SelfForcingFlowMatchScheduler":
+    ("schedulers", "scheduling_self_forcing_flow_match",
+     "SelfForcingFlowMatchScheduler"),
 }
 
 _FAST_VIDEO_MODELS = {

fastvideo/models/schedulers/scheduling_flow_match_euler_discrete.py

@@ -635,8 +635,31 @@ class FlowMatchEulerDiscreteScheduler(SchedulerMixin, ConfigMixin,
         noise: torch.Tensor,
         timestep: torch.IntTensor,
     ) -> torch.Tensor:
+
+        """
+        Args:
+            clean_latent: the clean latent with shape [B, C, H, W],
+                where B is batch_size or batch_size * num_frames
+            noise: the noise with shape [B, C, H, W]
+            timestep: the timestep with shape [1] or [bs * num_frames] or [bs, num_frames]
+
+        Returns:
+            the corrupted latent with shape [B, C, H, W]
+        """
+        # If timestep is [bs, num_frames]
+        if timestep.ndim == 2:
+            timestep = timestep.flatten(0, 1)
+            assert timestep.numel() == clean_latent.shape[0]
+        elif timestep.ndim == 1:
+            # If timestep is [1]
+            if timestep.shape[0] == 1:
+                timestep = timestep.expand(clean_latent.shape[0])
+            else:
+                assert timestep.numel() == clean_latent.shape[0]
+        else:
+            raise ValueError(f"[add_noise] Invalid timestep shape: {timestep.shape}")
+        # timestep shape should be [B]
         self.sigmas = self.sigmas.to(noise.device)
-        timestep = timestep.expand(clean_latent.shape[0])
         self.timesteps = self.timesteps.to(noise.device)
         timestep_id = torch.argmin(
             (self.timesteps.unsqueeze(0) - timestep.unsqueeze(1)).abs(), dim=1)

fastvideo/models/schedulers/scheduling_self_forcing_flow_match.py

@@ -0,0 +1,124 @@
+# SPDX-License-Identifier: Apache-2.0
+from diffusers.configuration_utils import ConfigMixin, register_to_config
+from diffusers.schedulers.scheduling_utils import SchedulerMixin
+from diffusers.utils import BaseOutput
+import torch
+
+from fastvideo.logger import init_logger
+from fastvideo.models.schedulers.base import BaseScheduler
+
+logger = init_logger(__name__)
+
+
+class SelfForcingFlowMatchSchedulerOutput(BaseOutput):
+    """
+    Output class for the scheduler's `step` function output.
+
+    Args:
+        prev_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
+            Computed sample `(x_{t-1})` of previous timestep. `prev_sample` should be used as next model input in the
+            denoising loop.
+    """
+    prev_sample: torch.FloatTensor
+
+class SelfForcingFlowMatchScheduler(BaseScheduler, ConfigMixin, SchedulerMixin):
+    
+    config_name = "scheduler_config.json"
+    order = 1
+    @register_to_config
+    def __init__(self, num_inference_steps=100, num_train_timesteps=1000, shift=3.0, sigma_max=1.0, sigma_min=0.003 / 1.002, inverse_timesteps=False, extra_one_step=False, reverse_sigmas=False, training=False):
+        self.num_train_timesteps = num_train_timesteps
+        self.shift = shift
+        self.sigma_max = sigma_max
+        self.sigma_min = sigma_min
+        self.inverse_timesteps = inverse_timesteps
+        self.extra_one_step = extra_one_step
+        self.reverse_sigmas = reverse_sigmas
+        self.set_timesteps(num_inference_steps, training=training)
+
+    def set_timesteps(self, num_inference_steps=100, denoising_strength=1.0, training=False, return_dict=False, **kwargs):
+        sigma_start = self.sigma_min + \
+            (self.sigma_max - self.sigma_min) * denoising_strength
+        if self.extra_one_step:
+            self.sigmas = torch.linspace(
+                sigma_start, self.sigma_min, num_inference_steps + 1)[:-1]
+        else:
+            self.sigmas = torch.linspace(
+                sigma_start, self.sigma_min, num_inference_steps)
+        if self.inverse_timesteps:
+            self.sigmas = torch.flip(self.sigmas, dims=[0])
+        self.sigmas = self.shift * self.sigmas / \
+            (1 + (self.shift - 1) * self.sigmas)
+        if self.reverse_sigmas:
+            self.sigmas = 1 - self.sigmas
+        self.timesteps = self.sigmas * self.num_train_timesteps
+        if training:
+            x = self.timesteps
+            y = torch.exp(-2 * ((x - num_inference_steps / 2) /
+                          num_inference_steps) ** 2)
+            y_shifted = y - y.min()
+            bsmntw_weighing = y_shifted * \
+                (num_inference_steps / y_shifted.sum())
+            self.linear_timesteps_weights = bsmntw_weighing
+
+    def step(self, model_output: torch.FloatTensor, timestep: torch.FloatTensor, sample: torch.FloatTensor, to_final=False, return_dict=False, **kwargs):
+        if timestep.ndim == 2:
+            timestep = timestep.flatten(0, 1)
+        self.sigmas = self.sigmas.to(model_output.device)
+        self.timesteps = self.timesteps.to(model_output.device)
+        timestep_id = torch.argmin(
+            (self.timesteps.unsqueeze(0) - timestep.unsqueeze(1)).abs(), dim=1)
+        sigma = self.sigmas[timestep_id].reshape(-1, 1, 1, 1)
+        if to_final or (timestep_id + 1 >= len(self.timesteps)).any():
+            sigma_ = 1 if (
+                self.inverse_timesteps or self.reverse_sigmas) else 0
+        else:
+            sigma_ = self.sigmas[timestep_id + 1].reshape(-1, 1, 1, 1)
+        prev_sample = sample + model_output * (sigma_ - sigma)
+        if isinstance(prev_sample, torch.Tensor | float) and not return_dict:
+            return (prev_sample, )
+        return SelfForcingFlowMatchSchedulerOutput(prev_sample=prev_sample)
+
+    def add_noise(self, original_samples, noise, timestep):
+        """
+        Diffusion forward corruption process.
+        Input:
+            - clean_latent: the clean latent with shape [B*T, C, H, W]
+            - noise: the noise with shape [B*T, C, H, W]
+            - timestep: the timestep with shape [B*T]
+        Output: the corrupted latent with shape [B*T, C, H, W]
+        """
+        if timestep.ndim == 2:
+            timestep = timestep.flatten(0, 1)
+        self.sigmas = self.sigmas.to(noise.device)
+        self.timesteps = self.timesteps.to(noise.device)
+        timestep_id = torch.argmin(
+            (self.timesteps.unsqueeze(0) - timestep.unsqueeze(1)).abs(), dim=1)
+        sigma = self.sigmas[timestep_id].reshape(-1, 1, 1, 1)
+        sample = (1 - sigma) * original_samples + sigma * noise
+        return sample.type_as(noise)
+
+    def training_target(self, sample, noise, timestep):
+        target = noise - sample
+        return target
+
+    def training_weight(self, timestep):
+        """
+        Input:
+            - timestep: the timestep with shape [B*T]
+        Output: the corresponding weighting [B*T]
+        """
+        if timestep.ndim == 2:
+            timestep = timestep.flatten(0, 1)
+        self.linear_timesteps_weights = self.linear_timesteps_weights.to(timestep.device)
+        timestep_id = torch.argmin(
+            (self.timesteps.unsqueeze(1) - timestep.unsqueeze(0)).abs(), dim=0)
+        weights = self.linear_timesteps_weights[timestep_id]
+        return weights
+    
+    def scale_model_input(self, sample: torch.Tensor, timestep: int | None = None) -> torch.Tensor:
+        return sample
+
+    def set_shift(self, shift: float) -> None:
+        self.shift = shift
+    

fastvideo/models/utils.py

@@ -6,9 +6,6 @@ from typing import Any
 import torch
 
 
-from fastvideo.logger import init_logger
-logger = init_logger(__name__)
-
 # TODO(PY): move it elsewhere
 def auto_attributes(init_func):
     """
@@ -148,14 +145,39 @@ def pred_noise_to_pred_video(pred_noise: torch.Tensor,
                              scheduler: Any) -> torch.Tensor:
     """
     Convert predicted noise to clean latent.
+
+    Args:
+    pred_noise: the predicted noise with shape [B, C, H, W]
+        where B is batch_size or batch_size * num_frames
+    noise_input_latent: the noisy latent with shape [B, C, H, W],
+    timestep: the timestep with shape [1] or [bs * num_frames] or [bs, num_frames]
+    scheduler: the scheduler
+
+    Returns:
+        the predicted video with shape [B, C, H, W]
     """
-    timestep = timestep.expand(noise_input_latent.shape[0])
+    # If timestep is [bs, num_frames]
+    if timestep.ndim == 2:
+        timestep = timestep.flatten(0, 1)
+        assert timestep.numel() == noise_input_latent.shape[0]
+    elif timestep.ndim == 1:
+        # If timestep is [1]
+        if timestep.shape[0] == 1:
+            timestep = timestep.expand(noise_input_latent.shape[0])
+        else:
+            assert timestep.numel() == noise_input_latent.shape[0]
+    else:
+        raise ValueError(f"[pred_noise_to_pred_video] Invalid timestep shape: {timestep.shape}")
+    # timestep shape should be [B]
     dtype = pred_noise.dtype
     device = pred_noise.device
-    pred_noise = pred_noise.float().to(device)
-    noise_input_latent = noise_input_latent.float().to(device)
-    sigmas = scheduler.sigmas.float().to(device)
-    timesteps = scheduler.timesteps.float().to(device)
+
+    # Convert to double following Self-Forcing
+    # https://github.com/guandeh17/Self-Forcing/blob/main/utils/wan_wrapper.py#L184
+    pred_noise = pred_noise.double().to(device)
+    noise_input_latent = noise_input_latent.double().to(device)
+    sigmas = scheduler.sigmas.double().to(device)
+    timesteps = scheduler.timesteps.double().to(device)
     timestep_id = torch.argmin(
         (timesteps.unsqueeze(0) - timestep.unsqueeze(1)).abs(), dim=1)
     sigma_t = sigmas[timestep_id].reshape(-1, 1, 1, 1)

fastvideo/pipelines/basic/wan/wan_causal_dmd_pipeline.py

@@ -16,8 +16,7 @@ from fastvideo.pipelines.stages import (ConditioningStage, DecodingStage,
                                         CausalDMDDenosingStage,
                                         InputValidationStage,
                                         LatentPreparationStage,
-                                        TextEncodingStage,
-                                        TimestepPreparationStage)
+                                        TextEncodingStage)
 # isort: on
 
 logger = init_logger(__name__)
@@ -29,10 +28,6 @@ class WanCausalDMDPipeline(LoRAPipeline, ComposedPipelineBase):
         "text_encoder", "tokenizer", "vae", "transformer", "scheduler"
     ]
 
-    def initialize_pipeline(self, fastvideo_args: FastVideoArgs):
-        self.modules["scheduler"] = FlowMatchEulerDiscreteScheduler(
-            shift=fastvideo_args.pipeline_config.flow_shift)
-
     def create_pipeline_stages(self, fastvideo_args: FastVideoArgs) -> None:
         """Set up pipeline stages with proper dependency injection."""
 
@@ -48,10 +43,6 @@ class WanCausalDMDPipeline(LoRAPipeline, ComposedPipelineBase):
         self.add_stage(stage_name="conditioning_stage",
                        stage=ConditioningStage())
 
-        self.add_stage(stage_name="timestep_preparation_stage",
-                       stage=TimestepPreparationStage(
-                           scheduler=self.get_module("scheduler")))
-
         self.add_stage(stage_name="latent_preparation_stage",
                        stage=LatentPreparationStage(
                            scheduler=self.get_module("scheduler"),

fastvideo/pipelines/composed_pipeline_base.py

@@ -79,7 +79,7 @@ class ComposedPipelineBase(ABC):
             for name, module in self.modules.items():
                 if not isinstance(module, torch.nn.Module):
                     continue
-                if name == "transformer":
+                if "transformer" in name:
                     module.requires_grad_(True)
                 else:
                     module.requires_grad_(False)
@@ -258,20 +258,19 @@ class ComposedPipelineBase(ABC):
         # remove keys that are not pipeline modules
         model_index.pop("_class_name")
         model_index.pop("_diffusers_version")
-        # @TODO(Wei): Temporary hack
         if "boundary_ratio" in model_index and model_index[
                 "boundary_ratio"] is not None:
             logger.info(
                 "MoE pipeline detected. Adding transformer_2 to self.required_config_modules..."
             )
             self.required_config_modules.append("transformer_2")
-            if fastvideo_args.boundary_ratio is None:
-                logger.info(
-                    "MoE pipeline detected. Setting boundary ratio to %s",
-                    model_index["boundary_ratio"])
-                fastvideo_args.boundary_ratio = model_index["boundary_ratio"]
+            logger.info("MoE pipeline detected. Setting boundary ratio to %s",
+                        model_index["boundary_ratio"])
+            fastvideo_args.pipeline_config.dit_config.boundary_ratio = model_index[
+                "boundary_ratio"]
 
         model_index.pop("boundary_ratio", None)
+        # used by Wan2.2 ti2v
         model_index.pop("expand_timesteps", None)
 
         # some sanity checks
@@ -304,8 +303,8 @@ class ComposedPipelineBase(ABC):
                           architecture) in model_index.items():
             if transformers_or_diffusers is None:
                 logger.warning(
-                    "Module in model_index.json has null value, removing from required_config_modules"
-                )
+                    "Module %s in model_index.json has null value, removing from required_config_modules",
+                    module_name)
                 if module_name in self.required_config_modules:
                     self.required_config_modules.remove(module_name)
                 continue

fastvideo/pipelines/lora_pipeline.py

@@ -7,7 +7,7 @@ import torch
 import torch.distributed as dist
 import torch.nn as nn
 from safetensors.torch import load_file
-from torch.distributed.device_mesh import init_device_mesh
+from torch.distributed.device_mesh import DeviceMesh, init_device_mesh
 from torch.distributed.tensor import DTensor
 
 from fastvideo.distributed import get_local_torch_device
@@ -32,6 +32,7 @@ class LoRAPipeline(ComposedPipelineBase):
     cur_adapter_name: str = ""
     cur_adapter_path: str = ""
     lora_layers: dict[str, BaseLayerWithLoRA] = {}
+    lora_layers_critic: dict[str, BaseLayerWithLoRA] = {}
     fastvideo_args: FastVideoArgs | TrainingArgs
     exclude_lora_layers: list[str] = []
     device: torch.device = get_local_torch_device()
@@ -81,6 +82,17 @@ class LoRAPipeline(ComposedPipelineBase):
 
     def set_trainable(self) -> None:
 
+        def set_lora_grads(lora_layers: dict[str, BaseLayerWithLoRA],
+                           device_mesh: DeviceMesh):
+            for name, layer in lora_layers.items():
+                layer.lora_A.requires_grad_(True)
+                layer.lora_B.requires_grad_(True)
+                layer.base_layer.requires_grad_(False)
+                layer.lora_A = nn.Parameter(
+                    DTensor.from_local(layer.lora_A, device_mesh=device_mesh))
+                layer.lora_B = nn.Parameter(
+                    DTensor.from_local(layer.lora_B, device_mesh=device_mesh))
+
         is_lora_training = self.training_mode and getattr(
             self.fastvideo_args, "lora_training", False)
         if not is_lora_training:
@@ -88,18 +100,12 @@ class LoRAPipeline(ComposedPipelineBase):
             return
 
         self.modules["transformer"].requires_grad_(False)
+        if "fake_score_transformer" in self.modules:
+            self.modules["fake_score_transformer"].requires_grad_(False)
         device_mesh = init_device_mesh("cuda", (dist.get_world_size(), 1),
                                        mesh_dim_names=["fake", "replicate"])
-        for name, layer in self.lora_layers.items():
-            # Enable grads for lora weights only
-            # Must convert to DTensor for compatibility with other FSDP modules in grad calculation
-            layer.lora_A.requires_grad_(True)
-            layer.lora_B.requires_grad_(True)
-            layer.base_layer.requires_grad_(False)
-            layer.lora_A = nn.Parameter(
-                DTensor.from_local(layer.lora_A, device_mesh=device_mesh))
-            layer.lora_B = nn.Parameter(
-                DTensor.from_local(layer.lora_B, device_mesh=device_mesh))
+        set_lora_grads(self.lora_layers, device_mesh)
+        set_lora_grads(self.lora_layers_critic, device_mesh)
 
     def convert_to_lora_layers(self) -> None:
         """
@@ -131,6 +137,24 @@ class LoRAPipeline(ComposedPipelineBase):
                 converted_count += 1
         logger.info("Converted %d layers to LoRA layers", converted_count)
 
+        if "fake_score_transformer" in self.modules:
+            for name, layer in self.modules[
+                    "fake_score_transformer"].named_modules():
+                if not self.is_target_layer(name):
+                    continue
+                layer = get_lora_layer(layer,
+                                       lora_rank=self.lora_rank,
+                                       lora_alpha=self.lora_alpha,
+                                       training_mode=self.training_mode)
+                if layer is not None:
+                    self.lora_layers_critic[name] = layer
+                    replace_submodule(self.modules["fake_score_transformer"],
+                                      name, layer)
+                    converted_count += 1
+            logger.info(
+                "Converted %d layers to LoRA layers in the critic model",
+                converted_count)
+
     def set_lora_adapter(self,
                          lora_nickname: str,
                          lora_path: str | None = None):  # type: ignore
@@ -224,4 +248,4 @@ class LoRAPipeline(ComposedPipelineBase):
 
     def unmerge_lora_weights(self) -> None:
         for name, layer in self.lora_layers.items():
-            layer.unmerge_lora_weights()
+            layer.unmerge_lora_weights()

fastvideo/pipelines/pipeline_batch_info.py

@@ -129,6 +129,7 @@ class ForwardBatch:
     timesteps: torch.Tensor | None = None
     timestep: torch.Tensor | float | int | None = None
     step_index: int | None = None
+    boundary_ratio: float | None = None
 
     # Scheduler parameters
     num_inference_steps: int = 50
@@ -245,10 +246,11 @@ class TrainingBatch:
     fake_score_loss: float = 0.0
 
     dmd_latent_vis_dict: dict[str, Any] = field(default_factory=dict)
+    latent_vis_dict: dict[str, torch.Tensor] = field(default_factory=dict)
     fake_score_latent_vis_dict: dict[str, Any] = field(default_factory=dict)
 
 
 @dataclass
 class PreprocessBatch(ForwardBatch):
-    video_loader: list["VideoDecoder"] = field(default_factory=list)
+    video_loader: list["VideoDecoder"] | list[str] = field(default_factory=list)
     video_file_name: list[str] = field(default_factory=list)

fastvideo/pipelines/preprocess/preprocess_pipeline_base.py

@@ -1,7 +1,5 @@
 # SPDX-License-Identifier: Apache-2.0
-import multiprocessing
 import os
-from concurrent.futures import ProcessPoolExecutor
 from typing import Any
 
 import numpy as np
@@ -12,6 +10,8 @@ from torch.utils.data import DataLoader
 from tqdm import tqdm
 
 from fastvideo.dataset import getdataset
+from fastvideo.dataset.dataloader.parquet_io import (ParquetDatasetWriter,
+                                                     records_to_table)
 from fastvideo.dataset.preprocessing_datasets import PreprocessBatch
 from fastvideo.distributed import get_local_torch_device
 from fastvideo.fastvideo_args import FastVideoArgs
@@ -54,10 +54,14 @@ class BasePreprocessPipeline(ComposedPipelineBase):
         """Get additional features specific to the pipeline type. Override in subclasses."""
         return {}
 
-    def get_schema_fields(self) -> list[str]:
-        """Get the schema fields for the pipeline type. Override in subclasses."""
+    def get_pyarrow_schema(self) -> pa.Schema:
+        """Return the PyArrow schema for this pipeline. Must be overridden."""
         raise NotImplementedError
 
+    def get_schema_fields(self) -> list[str]:
+        """Get the schema fields for the pipeline type."""
+        return [f.name for f in self.get_pyarrow_schema()]
+
     def create_record_for_schema(self,
                                  preprocess_batch: PreprocessBatch,
                                  schema: pa.Schema,
@@ -400,166 +404,22 @@ class BasePreprocessPipeline(ComposedPipelineBase):
                 batch_data.append(record)
 
             if batch_data:
-                # Add progress bar for writing to Parquet dataset
                 write_pbar = tqdm(total=1,
                                   desc="Writing to Parquet dataset",
                                   unit="batch")
-                # Convert batch data to PyArrow arrays
-                arrays = []
-                for field in self.get_schema_fields():
-                    if field.endswith('_bytes'):
-                        arrays.append(
-                            pa.array([record[field] for record in batch_data],
-                                     type=pa.binary()))
-                    elif field.endswith('_shape'):
-                        arrays.append(
-                            pa.array([record[field] for record in batch_data],
-                                     type=pa.list_(pa.int32())))
-                    elif field in ['width', 'height', 'num_frames']:
-                        arrays.append(
-                            pa.array([record[field] for record in batch_data],
-                                     type=pa.int32()))
-                    elif field in ['duration_sec', 'fps']:
-                        arrays.append(
-                            pa.array([record[field] for record in batch_data],
-                                     type=pa.float32()))
-                    else:
-                        arrays.append(
-                            pa.array([record[field] for record in batch_data]))
-
-                table = pa.Table.from_arrays(arrays,
-                                             names=self.get_schema_fields())
+                table = records_to_table(batch_data, self.get_pyarrow_schema())
                 write_pbar.update(1)
                 write_pbar.close()
 
-                # Store the table in a list for later processing
-                if not hasattr(self, 'all_tables'):
-                    self.all_tables = []
-                self.all_tables.append(table)
-
+                if not hasattr(self, 'dataset_writer'):
+                    self.dataset_writer = ParquetDatasetWriter(
+                        out_dir=combined_parquet_dir,
+                        samples_per_file=args.samples_per_file,
+                    )
+                self.dataset_writer.append_table(table)
                 logger.info("Collected batch with %s samples", len(table))
 
             if num_processed_samples >= args.flush_frequency:
-                self._flush_tables(num_processed_samples, args,
-                                   combined_parquet_dir)
+                written = self.dataset_writer.flush()
+                logger.info("Flushed %s samples to parquet", written)
                 num_processed_samples = 0
-                self.all_tables = []
-
-    def _flush_tables(self, num_processed_samples: int, args,
-                      combined_parquet_dir: str):
-        """Flush collected tables to disk."""
-        assert hasattr(self, 'all_tables') and self.all_tables
-        print(f"Combining {len(self.all_tables)} batches...")
-        combined_table = pa.concat_tables(self.all_tables)
-        assert len(combined_table) == num_processed_samples
-        print(f"Total samples collected: {len(combined_table)}")
-
-        # Calculate total number of chunks needed, discarding remainder
-        total_chunks = max(num_processed_samples // args.samples_per_file, 1)
-
-        print(f"Fixed samples per parquet file: {args.samples_per_file}")
-        print(f"Total number of parquet files: {total_chunks}")
-        print(
-            f"Total samples to be processed: {total_chunks * args.samples_per_file} (discarding {num_processed_samples % args.samples_per_file} samples)"
-        )
-
-        # Split work among processes
-        num_workers = int(min(multiprocessing.cpu_count(), total_chunks))
-        chunks_per_worker = (total_chunks + num_workers - 1) // num_workers
-
-        print(f"Using {num_workers} workers to process {total_chunks} chunks")
-        logger.info("Chunks per worker: %s", chunks_per_worker)
-
-        # Prepare work ranges
-        work_ranges = []
-        for i in range(num_workers):
-            start_idx = i * chunks_per_worker
-            end_idx = min((i + 1) * chunks_per_worker, total_chunks)
-            if start_idx < total_chunks:
-                work_ranges.append(
-                    (start_idx, end_idx, combined_table, i,
-                     combined_parquet_dir, args.samples_per_file))
-
-        total_written = 0
-        failed_ranges = []
-        with ProcessPoolExecutor(max_workers=num_workers) as executor:
-            futures = {
-                executor.submit(self.process_chunk_range, work_range):
-                work_range
-                for work_range in work_ranges
-            }
-            for future in tqdm(futures, desc="Processing chunks"):
-                try:
-                    written = future.result()
-                    total_written += written
-                    logger.info("Processed chunk with %s samples", written)
-                except Exception as e:
-                    work_range = futures[future]
-                    failed_ranges.append(work_range)
-                    logger.error("Failed to process range %s-%s: %s",
-                                 work_range[0], work_range[1], str(e))
-
-        # Retry failed ranges sequentially
-        if failed_ranges:
-            logger.warning("Retrying %s failed ranges sequentially",
-                           len(failed_ranges))
-            for work_range in failed_ranges:
-                try:
-                    total_written += self.process_chunk_range(work_range)
-                except Exception as e:
-                    logger.error(
-                        "Failed to process range %s-%s after retry: %s",
-                        work_range[0], work_range[1], str(e))
-
-        logger.info("Total samples written: %s", total_written)
-
-    @staticmethod
-    def process_chunk_range(args: Any) -> int:
-        start_idx, end_idx, table, worker_id, output_dir, samples_per_file = args
-        try:
-            total_written = 0
-            num_samples = len(table)
-
-            # Create worker-specific subdirectory
-            worker_dir = os.path.join(output_dir, f"worker_{worker_id}")
-            os.makedirs(worker_dir, exist_ok=True)
-
-            # Check how many files there are already in the dir, and update i accordingly
-            num_parquets = 0
-            for root, _, files in os.walk(worker_dir):
-                for file in files:
-                    if file.endswith('.parquet'):
-                        num_parquets += 1
-
-            for i in range(start_idx, end_idx):
-                start_sample = i * samples_per_file
-                end_sample = min((i + 1) * samples_per_file, num_samples)
-                chunk = table.slice(start_sample, end_sample - start_sample)
-
-                # Create chunk file in worker's directory
-                chunk_path = os.path.join(
-                    worker_dir, f"data_chunk_{i + num_parquets}.parquet")
-                temp_path = chunk_path + '.tmp'
-
-                try:
-                    # Write to temporary file
-                    pq.write_table(chunk, temp_path, compression='zstd')
-
-                    # Rename temporary file to final file
-                    if os.path.exists(chunk_path):
-                        os.remove(
-                            chunk_path)  # Remove existing file if it exists
-                    os.rename(temp_path, chunk_path)
-
-                    total_written += len(chunk)
-                except Exception as e:
-                    # Clean up temporary file if it exists
-                    if os.path.exists(temp_path):
-                        os.remove(temp_path)
-                    raise e
-
-            return total_written
-        except Exception as e:
-            logger.error("Error processing chunks %s-%s for worker %s: %s",
-                         start_idx, end_idx, worker_id, str(e))
-            raise

fastvideo/pipelines/preprocess/preprocess_pipeline_i2v.py

@@ -40,9 +40,9 @@ class PreprocessPipeline_I2V(BasePreprocessPipeline):
                            image_processor=self.get_module("image_processor"),
                        ))
 
-    def get_schema_fields(self) -> list[str]:
-        """Get the schema fields for I2V pipeline."""
-        return [f.name for f in pyarrow_schema_i2v]
+    def get_pyarrow_schema(self):
+        """Return the PyArrow schema for I2V pipeline."""
+        return pyarrow_schema_i2v
 
     def get_extra_features(self, valid_data: dict[str, Any],
                            fastvideo_args: FastVideoArgs) -> dict[str, Any]:

fastvideo/pipelines/preprocess/preprocess_pipeline_t2v.py

@@ -15,9 +15,9 @@ class PreprocessPipeline_T2V(BasePreprocessPipeline):
 
     _required_config_modules = ["text_encoder", "tokenizer", "vae"]
 
-    def get_schema_fields(self):
-        """Get the schema fields for T2V pipeline."""
-        return [f.name for f in pyarrow_schema_t2v]
+    def get_pyarrow_schema(self):
+        """Return the PyArrow schema for T2V pipeline."""
+        return pyarrow_schema_t2v
 
 
 EntryClass = PreprocessPipeline_T2V

fastvideo/pipelines/preprocess/preprocess_pipeline_text.py

@@ -0,0 +1,184 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+Text-only Data Preprocessing pipeline implementation.
+
+This module contains an implementation of the Text-only Data Preprocessing pipeline
+using the modular pipeline architecture, based on the ODE Trajectory preprocessing.
+"""
+
+import os
+from collections.abc import Iterator
+from typing import Any
+
+import torch
+from torch.utils.data import DataLoader
+from torchdata.stateful_dataloader import StatefulDataLoader
+from tqdm import tqdm
+
+from fastvideo.dataset import gettextdataset
+from fastvideo.dataset.dataloader.parquet_io import (ParquetDatasetWriter,
+                                                     records_to_table)
+from fastvideo.dataset.dataloader.record_schema import text_only_record_creator
+from fastvideo.dataset.dataloader.schema import pyarrow_schema_text_only
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.preprocess.preprocess_pipeline_base import (
+    BasePreprocessPipeline)
+from fastvideo.pipelines.stages import TextEncodingStage
+
+logger = init_logger(__name__)
+
+
+class PreprocessPipeline_Text(BasePreprocessPipeline):
+    """Text-only preprocessing pipeline implementation."""
+
+    _required_config_modules = ["text_encoder", "tokenizer"]
+    preprocess_dataloader: StatefulDataLoader
+    preprocess_loader_iter: Iterator[dict[str, Any]]
+    pbar: Any
+    num_processed_samples: int = 0
+
+    def get_pyarrow_schema(self):
+        """Return the PyArrow schema for text-only pipeline."""
+        return pyarrow_schema_text_only
+
+    def create_pipeline_stages(self, fastvideo_args: FastVideoArgs):
+        """Set up pipeline stages with proper dependency injection."""
+        self.add_stage(stage_name="prompt_encoding_stage",
+                       stage=TextEncodingStage(
+                           text_encoders=[self.get_module("text_encoder")],
+                           tokenizers=[self.get_module("tokenizer")],
+                       ))
+
+    def preprocess_text_only(self, fastvideo_args: FastVideoArgs, args):
+        """Preprocess text-only data."""
+
+        for batch_idx, data in enumerate(self.pbar):
+            if data is None:
+                continue
+
+            with torch.inference_mode():
+                # For text-only processing, we only need text data
+                # Filter out samples without text
+                valid_indices = []
+                for i, text in enumerate(data["text"]):
+                    if text and text.strip():  # Check if text is not empty
+                        valid_indices.append(i)
+                self.num_processed_samples += len(valid_indices)
+
+                if not valid_indices:
+                    continue
+
+                # Create new batch with only valid samples (text-only)
+                valid_data = {
+                    "text": [data["text"][i] for i in valid_indices],
+                    "path": [data["path"][i] for i in valid_indices],
+                }
+
+                batch_captions = valid_data["text"]
+                # Encode text using the standalone TextEncodingStage API
+                prompt_embeds_list, prompt_masks_list = self.prompt_encoding_stage.encode_text(
+                    batch_captions,
+                    fastvideo_args,
+                    encoder_index=[0],
+                    return_attention_mask=True,
+                )
+                prompt_embeds = prompt_embeds_list[0]
+                prompt_attention_masks = prompt_masks_list[0]
+                assert prompt_embeds.shape[0] == prompt_attention_masks.shape[0]
+
+                logger.info("===== prompt_embeds: %s", prompt_embeds.shape)
+                logger.info("===== prompt_attention_masks: %s",
+                            prompt_attention_masks.shape)
+
+                # Prepare batch data for Parquet dataset
+                batch_data = []
+
+                # Add progress bar for saving outputs
+                save_pbar = tqdm(enumerate(valid_data["path"]),
+                                 desc="Saving outputs",
+                                 unit="item",
+                                 leave=False)
+
+                for idx, text_path in save_pbar:
+                    text_name = os.path.basename(text_path).split(".")[0]
+
+                    # Convert tensors to numpy arrays
+                    text_embedding = prompt_embeds[idx].cpu().numpy()
+
+                    # Create record for Parquet dataset (text-only schema)
+                    record = text_only_record_creator(
+                        text_name=text_name,
+                        text_embedding=text_embedding,
+                        caption=valid_data["text"][idx],
+                    )
+                    batch_data.append(record)
+
+                if batch_data:
+                    write_pbar = tqdm(total=1,
+                                      desc="Writing to Parquet dataset",
+                                      unit="batch")
+                    table = records_to_table(batch_data,
+                                             pyarrow_schema_text_only)
+                    write_pbar.update(1)
+                    write_pbar.close()
+
+                    if not hasattr(self, 'dataset_writer'):
+                        self.dataset_writer = ParquetDatasetWriter(
+                            out_dir=self.combined_parquet_dir,
+                            samples_per_file=args.samples_per_file,
+                        )
+                    self.dataset_writer.append_table(table)
+
+                    logger.info("Collected batch with %s samples", len(table))
+
+                if self.num_processed_samples >= args.flush_frequency:
+                    written = self.dataset_writer.flush()
+                    logger.info("Flushed %s samples to parquet", written)
+                    self.num_processed_samples = 0
+
+        # Final flush for any remaining samples
+        if hasattr(self, 'dataset_writer'):
+            written = self.dataset_writer.flush(write_remainder=True)
+            if written:
+                logger.info("Final flush wrote %s samples", written)
+
+    # Text-only record creation moved to fastvideo.dataset.dataloader.record_schema
+
+    def forward(self, batch: ForwardBatch, fastvideo_args: FastVideoArgs, args):
+        if not self.post_init_called:
+            self.post_init()
+
+        self.local_rank = int(os.getenv("RANK", 0))
+        os.makedirs(args.output_dir, exist_ok=True)
+        # Create directory for combined data
+        self.combined_parquet_dir = os.path.join(args.output_dir,
+                                                 "combined_parquet_dataset")
+        os.makedirs(self.combined_parquet_dir, exist_ok=True)
+
+        # Loading text dataset
+        train_dataset = gettextdataset(args)
+
+        self.preprocess_dataloader = DataLoader(
+            train_dataset,
+            batch_size=args.preprocess_video_batch_size,
+            num_workers=args.dataloader_num_workers,
+        )
+
+        self.preprocess_loader_iter = iter(self.preprocess_dataloader)
+
+        self.num_processed_samples = 0
+        # Add progress bar for text preprocessing
+        self.pbar = tqdm(self.preprocess_loader_iter,
+                         desc="Processing text",
+                         unit="batch",
+                         disable=self.local_rank != 0)
+
+        # Initialize class variables for data sharing
+        self.text_data: dict[str, Any] = {}  # Store text metadata and paths
+
+        self.preprocess_text_only(fastvideo_args, args)
+
+
+EntryClass = PreprocessPipeline_Text

fastvideo/pipelines/preprocess/preprocess_stages.py

@@ -4,9 +4,11 @@ from typing import cast
 
 import numpy as np
 import torch
+import torchvision
 from einops import rearrange
 from torchvision import transforms
 
+from fastvideo.configs.configs import VideoLoaderType
 from fastvideo.dataset.transform import (CenterCropResizeVideo,
                                          TemporalRandomCrop)
 from fastvideo.fastvideo_args import FastVideoArgs, WorkloadType
@@ -61,7 +63,16 @@ class VideoTransformStage(PipelineStage):
                 else:
                     frame_indices = frame_indices[:self.num_frames]
 
-            video = batch.video_loader[i].get_frames_at(frame_indices).data
+            if fastvideo_args.preprocess_config.video_loader_type == VideoLoaderType.TORCHCODEC:
+                video = batch.video_loader[i].get_frames_at(frame_indices).data
+            elif fastvideo_args.preprocess_config.video_loader_type == VideoLoaderType.TORCHVISION:
+                video, _, _ = torchvision.io.read_video(batch.video_loader[i],
+                                                        output_format="TCHW")
+                video = video[frame_indices]
+            else:
+                raise ValueError(
+                    f"Invalid video loader type: {fastvideo_args.preprocess_config.video_loader_type}"
+                )
             video = self.video_transform(video)
             video_pixel_batch.append(video)
 

fastvideo/pipelines/preprocess/v1_preprocess.py

@@ -1,5 +1,6 @@
 import argparse
 import os
+from typing import Any
 
 from fastvideo import PipelineConfig
 from fastvideo.configs.models.vaes import WanVAEConfig
@@ -13,6 +14,8 @@ from fastvideo.pipelines.preprocess.preprocess_pipeline_ode_trajectory import (
     PreprocessPipeline_ODE_Trajectory)
 from fastvideo.pipelines.preprocess.preprocess_pipeline_t2v import (
     PreprocessPipeline_T2V)
+from fastvideo.pipelines.preprocess.preprocess_pipeline_text import (
+    PreprocessPipeline_Text)
 from fastvideo.utils import maybe_download_model
 
 logger = init_logger(__name__)
@@ -23,13 +26,22 @@ def main(args) -> None:
     maybe_init_distributed_environment_and_model_parallel(1, 1)
     num_gpus = int(os.environ["WORLD_SIZE"])
     assert num_gpus == 1, "Only support 1 GPU"
+
     pipeline_config = PipelineConfig.from_pretrained(args.model_path)
-    kwargs = {
-        "vae_precision": "fp32",
-        "vae_config": WanVAEConfig(load_encoder=True, load_decoder=True),
-        "flow_shift": 5,
-    }
+
+    kwargs: dict[str, Any] = {}
+    if args.preprocess_task == "text_only":
+        kwargs = {
+            "text_encoder_cpu_offload": False,
+        }
+    else:
+        # Full config for video/image processing
+        kwargs = {
+            "vae_precision": "fp32",
+            "vae_config": WanVAEConfig(load_encoder=True, load_decoder=True),
+        }
     pipeline_config.update_config_from_dict(kwargs)
+
     fastvideo_args = FastVideoArgs(
         model_path=args.model_path,
         num_gpus=get_world_size(),
@@ -42,14 +54,14 @@ def main(args) -> None:
         PreprocessPipeline = PreprocessPipeline_T2V
     elif args.preprocess_task == "i2v":
         PreprocessPipeline = PreprocessPipeline_I2V
-    elif args.preprocess_task == "ode_trajectory":
-        PreprocessPipeline = PreprocessPipeline_ODE_Trajectory
+    elif args.preprocess_task == "text_only":
+        PreprocessPipeline = PreprocessPipeline_Text
     else:
-        raise ValueError(f"Invalid preprocess task: {args.preprocess_task}")
+        raise ValueError(f"Invalid preprocess task: {args.preprocess_task}. "
+                         f"Valid options: t2v, i2v, ode_trajectory, text_only")
 
-    logger.info(
-        f"Preprocess task: {args.preprocess_task} using {PreprocessPipeline.__name__}"
-    )
+    logger.info("Preprocess task: %s using %s", args.preprocess_task,
+                PreprocessPipeline.__name__)
 
     pipeline = PreprocessPipeline(args.model_path, fastvideo_args)
     pipeline.forward(batch=None, fastvideo_args=fastvideo_args, args=args)
@@ -89,7 +101,11 @@ if __name__ == "__main__":
     parser.add_argument("--video_length_tolerance_range", type=int, default=2.0)
     parser.add_argument("--group_frame", action="store_true")  # TODO
     parser.add_argument("--group_resolution", action="store_true")  # TODO
-    parser.add_argument("--preprocess_task", type=str, default="t2v")
+    parser.add_argument("--preprocess_task",
+                        type=str,
+                        default="t2v",
+                        choices=["t2v", "i2v", "text_only"],
+                        help="Type of preprocessing task to run")
     parser.add_argument("--train_fps", type=int, default=30)
     parser.add_argument("--use_image_num", type=int, default=0)
     parser.add_argument("--text_max_length", type=int, default=256)

fastvideo/pipelines/stages/causal_denoising.py

@@ -4,11 +4,11 @@ from fastvideo.distributed import get_local_torch_device
 from fastvideo.fastvideo_args import FastVideoArgs
 from fastvideo.forward_context import set_forward_context
 from fastvideo.logger import init_logger
-from fastvideo.models.schedulers.scheduling_flow_match_euler_discrete import (
-    FlowMatchEulerDiscreteScheduler)
 from fastvideo.models.utils import pred_noise_to_pred_video
 from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
 from fastvideo.pipelines.stages.denoising import DenoisingStage
+from fastvideo.pipelines.stages.validators import StageValidators as V
+from fastvideo.pipelines.stages.validators import VerificationResult
 
 try:
     from fastvideo.attention.backends.sliding_tile_attn import (
@@ -36,7 +36,6 @@ class CausalDMDDenosingStage(DenoisingStage):
 
     def __init__(self, transformer, scheduler) -> None:
         super().__init__(transformer, scheduler)
-        self.scheduler = FlowMatchEulerDiscreteScheduler(shift=8.0)
         # KV and cross-attention cache state (initialized on first forward)
         self.kv_cache1: list | None = None
         self.crossattn_cache: list | None = None
@@ -71,8 +70,18 @@ class CausalDMDDenosingStage(DenoisingStage):
         # Timesteps for DMD
         timesteps = torch.tensor(
             fastvideo_args.pipeline_config.dmd_denoising_steps,
-            dtype=torch.long,
-            device=get_local_torch_device())
+            dtype=torch.long).cpu()
+
+        if fastvideo_args.pipeline_config.warp_denoising_step:
+            logger.info("Warping timesteps...")
+            scheduler_timesteps = torch.cat((self.scheduler.timesteps.cpu(),
+                                             torch.tensor([0],
+                                                          dtype=torch.float32)))
+            timesteps = scheduler_timesteps[1000 - timesteps]
+        else:
+            assert False, "warp_denoising_step must be true"
+        timesteps = timesteps.to(get_local_torch_device())
+        logger.info("Using timesteps: %s", timesteps)
 
         # Image kwargs (kept empty unless caller provides compatible args)
         image_kwargs: dict = {}
@@ -228,10 +237,7 @@ class CausalDMDDenosingStage(DenoisingStage):
                                                        dim=2)
 
                     # Prepare inputs
-                    t_expand = t_cur.expand(latent_model_input.shape[0])
-                    # t_expand = t_cur * torch.ones((latent_model_input.shape[0], 1), device=latent_model_input.device, dtype=torch.long)
-                    # t_expand = t_expand.repeat(1, self.sliding_window_num_frames)
-
+                    t_expand = t_cur.repeat(latent_model_input.shape[0])
 
                     # Attention metadata if needed
                     if (vsa_available and self.attn_backend
@@ -265,7 +271,10 @@ class CausalDMDDenosingStage(DenoisingStage):
                                             attn_metadata=attn_metadata,
                                             forward_batch=batch):
                         # Run transformer; follow DMD stage pattern
-                        t_expanded_noise= t_cur * torch.ones((latent_model_input.shape[0], 1), device=latent_model_input.device, dtype=torch.long)
+                        t_expanded_noise = t_cur * torch.ones(
+                            (latent_model_input.shape[0], 1),
+                            device=latent_model_input.device,
+                            dtype=torch.long)
                         pred_noise_btchw = self.transformer(
                             latent_model_input,
                             prompt_embeds,
@@ -330,7 +339,7 @@ class CausalDMDDenosingStage(DenoisingStage):
                     set_forward_context(current_timestep=0,
                                         attn_metadata=attn_metadata,
                                         forward_batch=batch):
-                    t_expanded_context = t_context * torch.ones((context_bcthw.shape[0], 1), device=context_bcthw.device, dtype=torch.long)
+                    t_expanded_context = t_context.unsqueeze(1)
                     _ = self.transformer(
                         context_bcthw,
                         prompt_embeds,
@@ -412,3 +421,27 @@ class CausalDMDDenosingStage(DenoisingStage):
                 False,
             })
         self.crossattn_cache = crossattn_cache
+
+    def verify_input(self, batch: ForwardBatch,
+                     fastvideo_args: FastVideoArgs) -> VerificationResult:
+        """Verify denoising stage inputs."""
+        result = VerificationResult()
+        result.add_check("latents", batch.latents,
+                         [V.is_tensor, V.with_dims(5)])
+        result.add_check("prompt_embeds", batch.prompt_embeds, V.list_not_empty)
+        result.add_check("image_embeds", batch.image_embeds, V.is_list)
+        result.add_check("image_latent", batch.image_latent,
+                         V.none_or_tensor_with_dims(5))
+        result.add_check("num_inference_steps", batch.num_inference_steps,
+                         V.positive_int)
+        result.add_check("guidance_scale", batch.guidance_scale,
+                         V.positive_float)
+        result.add_check("eta", batch.eta, V.non_negative_float)
+        result.add_check("generator", batch.generator,
+                         V.generator_or_list_generators)
+        result.add_check("do_classifier_free_guidance",
+                         batch.do_classifier_free_guidance, V.bool_value)
+        result.add_check(
+            "negative_prompt_embeds", batch.negative_prompt_embeds, lambda x:
+            not batch.do_classifier_free_guidance or V.list_not_empty(x))
+        return result

fastvideo/pipelines/stages/denoising.py

@@ -40,6 +40,13 @@ try:
 except ImportError:
     st_attn_available = False
 
+try:
+    from fastvideo.attention.backends.vmoba import VMOBAAttentionBackend
+    from fastvideo.utils import is_vmoba_available
+    vmoba_attn_available = is_vmoba_available()
+except ImportError:
+    vmoba_attn_available = False
+
 try:
     from fastvideo.attention.backends.video_sparse_attn import (
         VideoSparseAttentionBackend)
@@ -77,6 +84,7 @@ class DenoisingStage(PipelineStage):
             supported_attention_backends=(
                 AttentionBackendEnum.SLIDING_TILE_ATTN,
                 AttentionBackendEnum.VIDEO_SPARSE_ATTN,
+                AttentionBackendEnum.VMOBA_ATTN,
                 AttentionBackendEnum.FLASH_ATTN, AttentionBackendEnum.TORCH_SDPA
             )  # hack
         )
@@ -150,7 +158,8 @@ class DenoisingStage(PipelineStage):
         # Prepare image latents and embeddings for I2V generation
         image_embeds = batch.image_embeds
         if len(image_embeds) > 0:
-            assert torch.isnan(image_embeds[0]).sum() == 0
+            assert not torch.isnan(
+                image_embeds[0]).any(), "image_embeds contains nan"
             image_embeds = [
                 image_embed.to(target_dtype) for image_embed in image_embeds
             ]
@@ -187,15 +196,23 @@ class DenoisingStage(PipelineStage):
         # Get latents and embeddings
         latents = batch.latents
         prompt_embeds = batch.prompt_embeds
-        assert torch.isnan(prompt_embeds[0]).sum() == 0
+        assert not torch.isnan(
+            prompt_embeds[0]).any(), "prompt_embeds contains nan"
         if batch.do_classifier_free_guidance:
             neg_prompt_embeds = batch.negative_prompt_embeds
             assert neg_prompt_embeds is not None
-            assert torch.isnan(neg_prompt_embeds[0]).sum() == 0
+            assert not torch.isnan(
+                neg_prompt_embeds[0]).any(), "neg_prompt_embeds contains nan"
 
         # (Wan2.2) Calculate timestep to switch from high noise expert to low noise expert
-        if fastvideo_args.boundary_ratio is not None:
-            boundary_timestep = fastvideo_args.boundary_ratio * self.scheduler.num_train_timesteps
+        if fastvideo_args.pipeline_config.dit_config.boundary_ratio is not None:
+            boundary_timestep = fastvideo_args.pipeline_config.dit_config.boundary_ratio
+            if batch.boundary_timestep is not None:
+                logger.info("Overriding boundary timestep from %s to %s",
+                            boundary_timestep, batch.boundary_timestep)
+                boundary_timestep = batch.boundary_timestep
+
+            boundary_timestep *= self.scheduler.num_train_timesteps
         else:
             boundary_timestep = None
         latent_model_input = latents.to(target_dtype)
@@ -285,6 +302,9 @@ class DenoisingStage(PipelineStage):
                         logger.info("latent_model_input.shape: %s",
                                     latent_model_input.shape)
 
+                assert not torch.isnan(
+                    latent_model_input).any(), "latent_model_input contains nan"
+
                 if fastvideo_args.pipeline_config.ti2v_task and batch.pil_image is not None:
                     timestep = torch.stack([t]).to(get_local_torch_device())
                     temp_ts = (mask2[0][0][:, ::2, ::2] * timestep).flatten()
@@ -297,7 +317,6 @@ class DenoisingStage(PipelineStage):
                 else:
                     t_expand = t.repeat(latent_model_input.shape[0])
 
-                assert torch.isnan(latent_model_input).sum() == 0
                 latent_model_input = self.scheduler.scale_model_input(
                     latent_model_input, t)
                 if fastvideo_args.pipeline_config.t2v_as_i2v_task:
@@ -349,6 +368,31 @@ class DenoisingStage(PipelineStage):
                             assert attn_metadata is not None, "attn_metadata cannot be None"
                         else:
                             attn_metadata = None
+                    elif (vmoba_attn_available
+                          and self.attn_backend == VMOBAAttentionBackend):
+                        self.attn_metadata_builder_cls = self.attn_backend.get_builder_cls(
+                        )
+                        if self.attn_metadata_builder_cls is not None:
+                            self.attn_metadata_builder = self.attn_metadata_builder_cls(
+                            )
+                            # Prepare V-MoBA parameters from config
+                            moba_params = fastvideo_args.moba_config.copy()
+                            moba_params.update({
+                                "current_timestep":
+                                i,
+                                "raw_latent_shape":
+                                batch.raw_latent_shape[2:5],
+                                "patch_size":
+                                fastvideo_args.pipeline_config.dit_config.
+                                patch_size,
+                                "device":
+                                get_local_torch_device(),
+                            })
+                            attn_metadata = self.attn_metadata_builder.build(
+                                **moba_params)
+                            assert attn_metadata is not None, "attn_metadata cannot be None"
+                        else:
+                            attn_metadata = None
                     else:
                         attn_metadata = None
                     # TODO(will): finalize the interface. vLLM uses this to
@@ -791,7 +835,8 @@ class DmdDenoisingStage(DenoisingStage):
 
         video_raw_latent_shape = latents.shape
         prompt_embeds = batch.prompt_embeds
-        assert torch.isnan(prompt_embeds[0]).sum() == 0
+        assert not torch.isnan(
+            prompt_embeds[0]).any(), "prompt_embeds contains nan"
         timesteps = torch.tensor(
             fastvideo_args.pipeline_config.dmd_denoising_steps,
             dtype=torch.long,
@@ -830,7 +875,8 @@ class DmdDenoisingStage(DenoisingStage):
                         batch.image_latent.permute(0, 2, 1, 3, 4)
                     ],
                                                    dim=2).to(target_dtype)
-                assert torch.isnan(latent_model_input).sum() == 0
+                assert not torch.isnan(
+                    latent_model_input).any(), "latent_model_input contains nan"
 
                 # Prepare inputs for transformer
                 t_expand = t.repeat(latent_model_input.shape[0])

fastvideo/pipelines/stages/encoding.py

@@ -92,24 +92,6 @@ class EncodingStage(PipelineStage):
                 latents = latents.to(vae_dtype)
             latents = self.vae.encode(latents).mean
 
-        # Apply shifting if needed (reverse of decoding)
-        if (hasattr(self.vae, "shift_factor")
-                and self.vae.shift_factor is not None):
-            if isinstance(self.vae.shift_factor, torch.Tensor):
-                latents -= self.vae.shift_factor.to(latents.device,
-                                                    latents.dtype)
-            else:
-                latents -= self.vae.shift_factor
-
-        # Apply scaling factor
-        if (hasattr(self.vae, "scaling_factor")
-                and self.vae.scaling_factor is not None):
-            if isinstance(self.vae.scaling_factor, torch.Tensor):
-                latents = latents * self.vae.scaling_factor.to(
-                    latents.device, latents.dtype)
-            else:
-                latents = latents * self.vae.scaling_factor
-
         # Update batch with encoded latents
         batch.latents = latents
 

fastvideo/pipelines/stages/latent_preparation.py

@@ -59,10 +59,6 @@ class LatentPreparationStage(PipelineStage):
 
         # Adjust batch size for number of videos per prompt
         batch_size *= batch.num_videos_per_prompt
-        logger.info(f"===== batch_size: {batch_size}")
-        logger.info(f"===== batch.prompt: {batch.prompt}")
-        logger.info(f"===== batch.prompt_embeds: {batch.prompt_embeds}")
-        logger.info(f"===== batch.prompt_attention_mask: {batch.prompt_attention_mask}")
 
         # Get required parameters
         dtype = batch.prompt_embeds[0].dtype

fastvideo/platforms/cuda.py

@@ -159,6 +159,20 @@ class CudaPlatformBase(Platform):
                     str(e))
                 raise ImportError(
                     "Video Sparse Attention backend is not installed. ") from e
+        elif selected_backend == AttentionBackendEnum.VMOBA_ATTN:
+            try:
+                from csrc.attn.vmoba_attn.vmoba import (  # noqa: F401
+                    moba_attn_varlen)
+                from fastvideo.attention.backends.vmoba import (  # noqa: F401
+                    VMOBAAttentionBackend)
+                logger.info("Using Video MOBA Attention backend.")
+
+                return "fastvideo.attention.backends.vmoba.VMOBAAttentionBackend"
+            except ImportError as e:
+                logger.error(
+                    "Failed to import Video MoBA Attention backend: %s", str(e))
+                raise ImportError(
+                    "Video MoBA Attention backend is not installed. ") from e
         elif selected_backend == AttentionBackendEnum.TORCH_SDPA:
             logger.info("Using Torch SDPA backend.")
             return "fastvideo.attention.backends.sdpa.SDPABackend"

fastvideo/platforms/interface.py

@@ -19,6 +19,7 @@ class AttentionBackendEnum(enum.Enum):
     TORCH_SDPA = enum.auto()
     SAGE_ATTN = enum.auto()
     VIDEO_SPARSE_ATTN = enum.auto()
+    VMOBA_ATTN = enum.auto()
     NO_ATTENTION = enum.auto()
 
 

fastvideo/tests/dataset/test_parquet_io.py

@@ -0,0 +1,108 @@
+import os
+from pathlib import Path
+
+import pyarrow as pa
+import pyarrow.parquet as pq
+
+from fastvideo.dataset.dataloader.parquet_io import (
+    ParquetDatasetWriter,
+    records_to_table,
+)
+
+
+def test_records_to_table_types():
+    schema = pa.schema([
+        pa.field("id", pa.string()),
+        pa.field("vae_latent_bytes", pa.binary()),
+        pa.field("vae_latent_shape", pa.list_(pa.int64())),
+        pa.field("duration_sec", pa.float64()),
+        pa.field("width", pa.int64()),
+    ])
+    records = [{
+        "id": "a",
+        "vae_latent_bytes": b"\x00\x01",
+        "vae_latent_shape": [1, 2, 3],
+        "duration_sec": 1.5,
+        "width": 640,
+    }]
+
+    table = records_to_table(records, schema)
+    assert table.schema == schema
+    assert table.num_rows == 1
+    cols = {name: table.column(name).to_pylist()[0] for name in schema.names}
+    assert cols["id"] == "a"
+    assert isinstance(cols["vae_latent_bytes"], (bytes, bytearray))
+    assert cols["vae_latent_shape"] == [1, 2, 3]
+    assert abs(cols["duration_sec"] - 1.5) < 1e-6
+    assert cols["width"] == 640
+
+
+def test_writer_flush_and_remainder(tmp_path: Path):
+    schema = pa.schema([pa.field("id", pa.string())])
+    records = [{"id": str(i)} for i in range(25)]
+    table = records_to_table(records, schema)
+
+    out_dir = tmp_path / "out"
+    writer = ParquetDatasetWriter(str(out_dir), samples_per_file=10)
+    writer.append_table(table)
+    written = writer.flush(num_workers=1)
+    assert written == 20
+
+    files = sorted(out_dir.rglob("*.parquet"))
+    assert len(files) == 2
+    total_rows = sum(pq.read_table(str(f)).num_rows for f in files)
+    assert total_rows == 20
+
+    # Append remainder to complete another chunk
+    extra = records_to_table([{"id": str(i)} for i in range(5)], schema)
+    writer.append_table(extra)
+    written2 = writer.flush(num_workers=1)
+    assert written2 == 10
+    files2 = sorted(out_dir.rglob("*.parquet"))
+    assert len(files2) == 3
+    total_rows2 = sum(pq.read_table(str(f)).num_rows for f in files2)
+    assert total_rows2 == 30
+
+
+def test_writer_flush_write_remainder(tmp_path: Path):
+    schema = pa.schema([pa.field("id", pa.string())])
+    # 25 rows, 10 per file => 2 full files + 1 remainder(5)
+    records = [{"id": str(i)} for i in range(25)]
+    table = records_to_table(records, schema)
+
+    out_dir = tmp_path / "out_last"
+    writer = ParquetDatasetWriter(str(out_dir), samples_per_file=10)
+    writer.append_table(table)
+    # First flush writes 20
+    written1 = writer.flush(num_workers=1)
+    assert written1 == 20
+    # Final flush with remainder
+    written2 = writer.flush(num_workers=1, write_remainder=True)
+    assert written2 == 5
+    files = sorted(out_dir.rglob("*.parquet"))
+    assert len(files) == 3
+    total_rows = sum(pq.read_table(str(f)).num_rows for f in files)
+    assert total_rows == 25
+
+
+def test_writer_parallel_workers(tmp_path: Path):
+    schema = pa.schema([pa.field("id", pa.string())])
+    # 40 rows, 10 per file => 4 files
+    records = [{"id": str(i)} for i in range(40)]
+    table = records_to_table(records, schema)
+
+    out_dir = tmp_path / "out_parallel"
+    writer = ParquetDatasetWriter(str(out_dir), samples_per_file=10)
+    writer.append_table(table)
+    written = writer.flush(num_workers=2)
+    assert written == 40
+
+    # Ensure files exist under worker subdirs
+    worker_dirs = [p for p in out_dir.iterdir() if p.is_dir() and p.name.startswith("worker_")]
+    assert len(worker_dirs) >= 1
+    files = sorted(out_dir.rglob("*.parquet"))
+    assert len(files) == 4
+    total_rows = sum(pq.read_table(str(f)).num_rows for f in files)
+    assert total_rows == 40
+
+

fastvideo/tests/dataset/test_schema_record_creator.py

@@ -0,0 +1,123 @@
+import numpy as np
+
+from fastvideo.dataset.dataloader.record_schema import (
+    basic_t2v_record_creator,
+    i2v_record_creator,
+    ode_text_only_record_creator,
+    text_only_record_creator,
+)
+from fastvideo.pipelines.pipeline_batch_info import PreprocessBatch
+
+
+def _mk_basic_batch(N: int) -> PreprocessBatch:
+    batch = PreprocessBatch(data_type="video")
+    batch.video_file_name = [f"vid_{i}" for i in range(N)]
+    batch.prompt = [f"caption_{i}" for i in range(N)]
+    batch.width = [640 for _ in range(N)]
+    batch.height = [360 for _ in range(N)]
+    batch.fps = [4 for _ in range(N)]
+    batch.num_frames = [2 for _ in range(N)]
+    # Latents: shape (N, C, T, H, W); per-record use latents[idx]
+    batch.latents = np.zeros((N, 4, 2, 8, 8), dtype=np.float32)
+    # Prompt embeds: list of per-record arrays [Seq, Dim]
+    batch.prompt_embeds = [np.ones((6, 16), dtype=np.float32) for _ in range(N)]
+    return batch
+
+
+def test_basic_t2v_record_creator_fields():
+    N = 2
+    batch = _mk_basic_batch(N)
+
+    records = basic_t2v_record_creator(batch)
+    assert isinstance(records, list) and len(records) == N
+
+    for i, rec in enumerate(records):
+        assert rec["id"] == batch.video_file_name[i]
+        # Latents bytes/shape/dtype
+        assert isinstance(rec["vae_latent_bytes"], (bytes, bytearray))
+        assert rec["vae_latent_shape"] == list(batch.latents[i].shape)
+        assert rec["vae_latent_dtype"] == str(batch.latents[i].dtype)
+        # Text embedding
+        assert isinstance(rec["text_embedding_bytes"], (bytes, bytearray))
+        assert rec["text_embedding_shape"] == list(batch.prompt_embeds[i].shape)
+        assert rec["text_embedding_dtype"] == str(batch.prompt_embeds[i].dtype)
+        # Meta
+        assert rec["caption"] == batch.prompt[i]
+        assert rec["media_type"] == "video"
+        assert rec["width"] == int(batch.width[i])
+        assert rec["height"] == int(batch.height[i])
+        assert rec["num_frames"] == batch.latents[i].shape[1]
+
+
+def test_i2v_record_creator_additional_fields():
+    N = 3
+    batch = _mk_basic_batch(N)
+    # image_embeds is a list of length 1, with an array of shape [N, D]
+    batch.image_embeds = [np.ones((N, 32), dtype=np.float32)]
+    # first frame latent per record
+    batch.image_latent = np.zeros((N, 4, 1, 8, 8), dtype=np.float32)
+    # pil image per record
+    batch.pil_image = np.zeros((N, 8, 8, 3), dtype=np.uint8)
+
+    records = i2v_record_creator(batch)
+    assert isinstance(records, list) and len(records) == N
+
+    for i, rec in enumerate(records):
+        # clip feature
+        assert isinstance(rec["clip_feature_bytes"], (bytes, bytearray))
+        assert rec["clip_feature_shape"] == list(batch.image_embeds[0][i].shape)
+        assert rec["clip_feature_dtype"] == str(batch.image_embeds[0][i].dtype)
+        # first frame latent
+        assert isinstance(rec["first_frame_latent_bytes"], (bytes, bytearray))
+        assert rec["first_frame_latent_shape"] == list(batch.image_latent[i].shape)
+        assert rec["first_frame_latent_dtype"] == str(batch.image_latent[i].dtype)
+        # pil image
+        assert isinstance(rec["pil_image_bytes"], (bytes, bytearray))
+        assert rec["pil_image_shape"] == list(batch.pil_image[i].shape)
+        assert rec["pil_image_dtype"] == str(batch.pil_image[i].dtype)
+
+
+def test_ode_text_only_record_creator():
+    video_name = "ex"
+    caption = "a prompt"
+    text_embedding = np.ones((6, 16), dtype=np.float32)
+    traj = np.ones((5, 4, 2, 2), dtype=np.float32)
+    tsteps = np.arange(5, dtype=np.float32)
+
+    rec = ode_text_only_record_creator(
+        video_name=video_name,
+        text_embedding=text_embedding,
+        caption=caption,
+        trajectory_latents=traj,
+        trajectory_timesteps=tsteps,
+    )
+    assert rec["id"] == f"text_{video_name}"
+    assert isinstance(rec["text_embedding_bytes"], (bytes, bytearray))
+    assert rec["text_embedding_shape"] == list(text_embedding.shape)
+    assert rec["text_embedding_dtype"] == str(text_embedding.dtype)
+    assert rec["file_name"] == video_name
+    assert rec["caption"] == caption
+    assert rec["media_type"] == "text"
+    # Trajectory fields
+    assert isinstance(rec["trajectory_latents_bytes"], (bytes, bytearray))
+    assert rec["trajectory_latents_shape"] == list(traj.shape)
+    assert rec["trajectory_latents_dtype"] == str(traj.dtype)
+    assert isinstance(rec["trajectory_timesteps_bytes"], (bytes, bytearray))
+    assert rec["trajectory_timesteps_shape"] == list(tsteps.shape)
+    assert rec["trajectory_timesteps_dtype"] == str(tsteps.dtype)
+
+
+def test_text_only_record_creator():
+    text_name = "note1"
+    caption = "a prompt"
+    text_embedding = np.ones((7, 16), dtype=np.float32)
+    rec = text_only_record_creator(
+        text_name=text_name,
+        text_embedding=text_embedding,
+        caption=caption,
+    )
+    assert rec["id"] == f"text_{text_name}"
+    assert isinstance(rec["text_embedding_bytes"], (bytes, bytearray))
+    assert rec["text_embedding_shape"] == list(text_embedding.shape)
+    assert rec["text_embedding_dtype"] == str(text_embedding.dtype)
+    assert rec["caption"] == caption

fastvideo/tests/inference/vmoba/test_vmoba_inference.py

@@ -0,0 +1,58 @@
+# SPDX-License-Identifier: Apache-2.0
+
+import os
+import subprocess
+from pathlib import Path
+
+def test_inference_vmoba():
+    """Test FastVideo VMOBA_ATTN inference pipeline"""
+
+    num_gpus = "1"
+    model_base = "Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
+    output_dir = Path("outputs_video/vmoba_1.3B/")
+    moba_config = "fastvideo/configs/backend/vmoba/wan_1.3B_77_480_832.json"
+
+    os.environ["FASTVIDEO_ATTENTION_BACKEND"] = "VMOBA_ATTN"
+
+    cmd = [
+        "fastvideo", "generate",
+        "--model-path", model_base,
+        "--sp-size", num_gpus,
+        "--tp-size", "1",
+        "--num-gpus", num_gpus,
+        "--dit-cpu-offload", "False",
+        "--vae-cpu-offload", "False",
+        "--text-encoder-cpu-offload", "True",
+        "--pin-cpu-memory", "False",
+        "--height", "480",
+        "--width", "832",
+        "--num-frames", "77",
+        "--num-inference-steps", "50",
+        "--moba-config-path", moba_config,
+        "--fps", "16",
+        "--guidance-scale", "6.0",
+        "--flow-shift", "8.0",
+        "--prompt", "A majestic lion strides across the golden savanna, its powerful frame glistening under the warm afternoon sun. The tall grass ripples gently in the breeze, enhancing the lion's commanding presence. The tone is vibrant, embodying the raw energy of the wild. Low angle, steady tracking shot, cinematic.",
+        "--negative-prompt", (
+            "Bright tones, overexposed, static, blurred details, subtitles, style, "
+            "works, paintings, images, static, overall gray, worst quality, low quality, "
+            "JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, "
+            "poorly drawn faces, deformed, disfigured, misshapen limbs, fused fingers, "
+            "still picture, messy background, three legs, many people in the background, walking backwards"
+        ),
+        "--seed", "1024",
+        "--output-path", str(output_dir),
+    ]
+
+    subprocess.run(cmd, check=True)
+
+    assert output_dir.exists(), f"Output directory {output_dir} does not exist"
+
+    video_files = list(output_dir.glob("*.mp4"))
+    assert len(video_files) > 0, "No video files were generated"
+
+    for video_file in video_files:
+        assert video_file.stat().st_size > 0, f"Video file {video_file} is empty"
+
+if __name__ == "__main__":
+    test_inference_vmoba()

fastvideo/tests/modal/pr_test.py

@@ -102,10 +102,22 @@ def run_precision_tests_STA():
 def run_precision_tests_VSA():
     run_test("python csrc/attn/tests/test_vsa.py")
 
+@app.function(gpu="L40S:1", image=image, timeout=900)
+def run_precision_tests_vmoba():
+    run_test("pytest csrc/attn/vmoba_attn/tests/test_vmoba_attn.py")
+
+@app.function(gpu="L40S:1", image=image, timeout=900)
+def run_inference_tests_vmoba():
+    run_test('python fastvideo/tests/inference/vmoba/test_vmoba_inference.py')
+
 @app.function(gpu="L40S:1", image=image, timeout=3600)
 def run_inference_lora_tests():
     run_test("pytest ./fastvideo/tests/inference/lora/test_lora_inference_similarity.py -vs")
 
 @app.function(gpu="L40S:2", image=image, timeout=900)
 def run_distill_dmd_tests():
-    run_test("pytest ./fastvideo/tests/training/distill/test_distill_dmd.py -vs")
+    run_test("pytest ./fastvideo/tests/training/distill/test_distill_dmd.py -vs")
+
+@app.function(gpu="L40S:1", image=image, timeout=900)
+def run_unit_test():
+    run_test("pytest ./fastvideo/tests/dataset/ ./fastvideo/tests/workflow/ -vs")

fastvideo/tests/ssim/README.md

@@ -4,7 +4,8 @@ The reference videos in the `*_reference_videos` directory are used as part of a
 
 run `bash update_reference_videos.sh` from inside the `fastvideo/tests/ssim/` directory after running `test_inference_similarity.py` to update reference videos. Note: make sure to update the path to the corresponding device.
 
-all reference videos are were generated on commit `4aeabbc629e0edf91477e80e795e7bb1823c71cb`
+reference videos were generated on commit `4aeabbc629e0edf91477e80e795e7bb1823c71cb`
+causal videos were generated on commit b318063c0a4618f1d5d99ea82ca67a06aad0d19d
 
 ## Generation Details
 
@@ -76,4 +77,4 @@ Wan2.1-I2V-14B-480P-Diffusers: {
 ### Image-to-Video Prompts
 
 1. "An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space realised in the background. High quality, ultrarealistic detail and breath-taking movie-like camera shot."
-    Image path: "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/astronaut.jpg"
+    Image path: "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/astronaut.jpg"

fastvideo/tests/ssim/test_causal_similarity.py

@@ -0,0 +1,152 @@
+# SPDX-License-Identifier: Apache-2.0
+import json
+import os
+
+import torch
+import pytest
+
+from fastvideo import VideoGenerator
+from fastvideo.logger import init_logger
+from fastvideo.tests.utils import compute_video_ssim_torchvision, write_ssim_results
+from fastvideo.worker.multiproc_executor import MultiprocExecutor
+
+logger = init_logger(__name__)
+
+device_name = torch.cuda.get_device_name()
+device_reference_folder_suffix = '_reference_videos'
+
+if "A40" in device_name:
+    device_reference_folder = "A40" + device_reference_folder_suffix
+elif "L40S" in device_name:
+    device_reference_folder = "L40S" + device_reference_folder_suffix
+
+# Base parameters from the shell script
+
+SF_WAN_T2V_PARAMS = {
+    "num_gpus": 1,
+    "model_path": "wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers",
+    "height": 480,
+    "width": 832,
+    "num_frames": 81,
+    "num_inference_steps": 4,
+    "seed": 1024,
+    "sp_size": 1,
+    "tp_size": 1,
+}
+
+
+MODEL_TO_PARAMS = {
+    "SFWan2.1-T2V-1.3B-Diffusers": SF_WAN_T2V_PARAMS,
+}
+
+I2V_MODEL_TO_PARAMS = {
+}
+
+TEST_PROMPTS = [
+    "Will Smith casually eats noodles, his relaxed demeanor contrasting with the energetic background of a bustling street food market. The scene captures a mix of humor and authenticity. Mid-shot framing, vibrant lighting.",
+    # "A lone hiker stands atop a towering cliff, silhouetted against the vast horizon. The rugged landscape stretches endlessly beneath, its earthy tones blending into the soft blues of the sky. The scene captures the spirit of exploration and human resilience. High angle, dynamic framing, with soft natural lighting emphasizing the grandeur of nature."
+]
+
+I2V_TEST_PROMPTS = [
+    "An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space realised in the background. High quality, ultrarealistic detail and breath-taking movie-like camera shot.",
+]
+
+I2V_IMAGE_PATHS = [
+    "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/astronaut.jpg",
+]
+
+@pytest.mark.parametrize("prompt", TEST_PROMPTS)
+@pytest.mark.parametrize("ATTENTION_BACKEND", ["FLASH_ATTN"])
+@pytest.mark.parametrize("model_id", list(MODEL_TO_PARAMS.keys()))
+def test_causal_similarity(prompt, ATTENTION_BACKEND, model_id):
+    """
+    Test that runs inference with different parameters and compares the output
+    to reference videos using SSIM.
+    """
+    os.environ["FASTVIDEO_ATTENTION_BACKEND"] = ATTENTION_BACKEND
+
+    script_dir = os.path.dirname(os.path.abspath(__file__))
+
+    base_output_dir = os.path.join(script_dir, 'generated_videos', model_id)
+    output_dir = os.path.join(base_output_dir, ATTENTION_BACKEND)
+    output_video_name = f"{prompt[:100]}.mp4"
+
+    os.makedirs(output_dir, exist_ok=True)
+
+    BASE_PARAMS = MODEL_TO_PARAMS[model_id]
+    num_inference_steps = BASE_PARAMS["num_inference_steps"]
+
+    init_kwargs = {
+        "num_gpus": BASE_PARAMS["num_gpus"],
+        "sp_size": BASE_PARAMS["sp_size"],
+        "tp_size": BASE_PARAMS["tp_size"],
+        "dit_cpu_offload": True,
+    }
+    if BASE_PARAMS.get("vae_sp"):
+        init_kwargs["vae_sp"] = True
+        init_kwargs["vae_tiling"] = True
+    #if "text-encoder-precision" in BASE_PARAMS:
+    #    init_kwargs["text_encoder_precisions"] = BASE_PARAMS["text-encoder-precision"]
+
+    generation_kwargs = {
+        "num_inference_steps": num_inference_steps,
+        "output_path": output_dir,
+        "height": BASE_PARAMS["height"],
+        "width": BASE_PARAMS["width"],
+        "num_frames": BASE_PARAMS["num_frames"],
+        "seed": BASE_PARAMS["seed"],
+    }
+    if "neg_prompt" in BASE_PARAMS:
+        generation_kwargs["neg_prompt"] = BASE_PARAMS["neg_prompt"]
+
+    generator = VideoGenerator.from_pretrained(model_path=BASE_PARAMS["model_path"], **init_kwargs)
+    generator.generate_video(prompt, **generation_kwargs)
+
+    if isinstance(generator.executor, MultiprocExecutor):
+        generator.executor.shutdown()
+
+    assert os.path.exists(
+        output_dir), f"Output video was not generated at {output_dir}"
+
+    reference_folder = os.path.join(script_dir, device_reference_folder, model_id, ATTENTION_BACKEND)
+
+    if not os.path.exists(reference_folder):
+        logger.error("Reference folder missing")
+        raise FileNotFoundError(
+            f"Reference video folder does not exist: {reference_folder}")
+
+    # Find the matching reference video based on the prompt
+    reference_video_name = None
+
+    for filename in os.listdir(reference_folder):
+        if filename.endswith('.mp4') and prompt[:100] in filename:
+            reference_video_name = filename
+            break
+
+    if not reference_video_name:
+        logger.error(f"Reference video not found for prompt: {prompt} with backend: {ATTENTION_BACKEND}")
+        raise FileNotFoundError(f"Reference video missing")
+
+    reference_video_path = os.path.join(reference_folder, reference_video_name)
+    generated_video_path = os.path.join(output_dir, output_video_name)
+
+    logger.info(
+        f"Computing SSIM between {reference_video_path} and {generated_video_path}"
+    )
+    ssim_values = compute_video_ssim_torchvision(reference_video_path,
+                                                 generated_video_path,
+                                                 use_ms_ssim=True)
+
+    mean_ssim = ssim_values[0]
+    logger.info(f"SSIM mean value: {mean_ssim}")
+    logger.info(f"Writing SSIM results to directory: {output_dir}")
+
+    success = write_ssim_results(output_dir, ssim_values, reference_video_path,
+                                 generated_video_path, num_inference_steps,
+                                 prompt)
+
+    if not success:
+        logger.error("Failed to write SSIM results to file")
+
+    min_acceptable_ssim = 0.98
+    assert mean_ssim >= min_acceptable_ssim, f"SSIM value {mean_ssim} is below threshold {min_acceptable_ssim} for {model_id} with backend {ATTENTION_BACKEND}"

fastvideo/tests/ssim/test_inference_similarity.py

@@ -101,7 +101,7 @@ I2V_IMAGE_PATHS = [
 
 
 @pytest.mark.parametrize("prompt", I2V_TEST_PROMPTS)
-@pytest.mark.parametrize("ATTENTION_BACKEND", ["FLASH_ATTN", "TORCH_SDPA"])
+@pytest.mark.parametrize("ATTENTION_BACKEND", ["FLASH_ATTN"])
 @pytest.mark.parametrize("model_id", list(I2V_MODEL_TO_PARAMS.keys()))
 def test_i2v_inference_similarity(prompt, ATTENTION_BACKEND, model_id):
     """

fastvideo/tests/transformers/test_ori_causal_wanvideo.py

@@ -0,0 +1,292 @@
+# SPDX-License-Identifier: Apache-2.0
+import os
+import math
+
+import numpy as np
+import pytest
+import torch
+from fastvideo.wan.modules.causal_model import CausalWanModel
+
+from fastvideo.configs.pipelines import PipelineConfig
+from fastvideo.forward_context import set_forward_context
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.loader.component_loader import TransformerLoader
+from fastvideo.models.dits.causal_wanvideo import CausalWanTransformer3DModel
+from fastvideo.utils import maybe_download_model
+from fastvideo.configs.models.dits import WanVideoConfig
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+
+
+logger = init_logger(__name__)
+
+os.environ["MASTER_ADDR"] = "localhost"
+os.environ["MASTER_PORT"] = "29503"
+
+BASE_MODEL_PATH = "wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers"
+MODEL_PATH = maybe_download_model(BASE_MODEL_PATH,
+                                  local_dir=os.path.join(
+                                      'data', BASE_MODEL_PATH))
+TRANSFORMER_PATH = os.path.join(MODEL_PATH, "transformer")
+
+
+@pytest.mark.usefixtures("distributed_setup")
+def test_ori_causal_wan_transformer():
+    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+    precision = torch.bfloat16
+    precision_str = "bf16"
+    args = FastVideoArgs(model_path=TRANSFORMER_PATH,
+                         dit_cpu_offload=True,
+                         pipeline_config=PipelineConfig(dit_config=WanVideoConfig(), dit_precision=precision_str))
+    args.device = device
+
+    loader = TransformerLoader()
+    model2 = loader.load(TRANSFORMER_PATH, args).to(dtype=precision)
+
+    model1 = CausalWanModel.from_pretrained(
+        "/mnt/weka/home/hao.zhang/wei/Self-Forcing/wan_models/Wan2.1-T2V-1.3B", device=device,
+        torch_dtype=precision).to(device, dtype=precision).requires_grad_(False)
+    causal_state_dict = torch.load("/mnt/weka/home/hao.zhang/wei/Self-Forcing/checkpoints/self_forcing_dmd.pt")["generator_ema"]
+    new_state_dict = {}
+    for k, v in causal_state_dict.items():
+        if k.startswith("model."):
+            new_state_dict[k.replace("model.", "")] = v
+    causal_state_dict = new_state_dict
+    model1.load_state_dict(causal_state_dict)
+
+    total_params = sum(p.numel() for p in model1.parameters())
+    # Calculate weight sum for model1 (converting to float64 to avoid overflow)
+    weight_sum_model1 = sum(
+        p.to(torch.float64).sum().item() for p in model1.parameters())
+    # Also calculate mean for more stable comparison
+    weight_mean_model1 = weight_sum_model1 / total_params
+    logger.info("Model 1 weight sum: %s", weight_sum_model1)
+    logger.info("Model 1 weight mean: %s", weight_mean_model1)
+
+    # Calculate weight sum for model2 (converting to float64 to avoid overflow)
+    total_params_model2 = sum(p.numel() for p in model2.parameters())
+    weight_sum_model2 = sum(
+        p.to(torch.float64).sum().item() for p in model2.parameters())
+    # Also calculate mean for more stable comparison
+    weight_mean_model2 = weight_sum_model2 / total_params_model2
+    logger.info("Model 2 weight sum: %s", weight_sum_model2)
+    logger.info("Model 2 weight mean: %s", weight_mean_model2)
+
+    weight_sum_diff = abs(weight_sum_model1 - weight_sum_model2)
+    logger.info("Weight sum difference: %s", weight_sum_diff)
+    weight_mean_diff = abs(weight_mean_model1 - weight_mean_model2)
+    logger.info("Weight mean difference: %s", weight_mean_diff)
+
+    # Set both models to eval mode
+    model1 = model1.eval()
+    model2 = model2.eval()
+
+    # Create identical inputs for both models
+    batch_size = 1
+    text_seq_len = 30
+
+    # Video latents [B, C, T, H, W]
+    hidden_states = torch.randn(batch_size,
+                                16,
+                                12,
+                                160,
+                                90,
+                                device=device,
+                                dtype=precision)
+    block_sizes = [3 for _ in range(4)]
+    timesteps = [1000, 750, 500, 250]
+
+    # Text embeddings [B, L, D] (including global token)
+    encoder_hidden_states = torch.randn(batch_size,
+                                        text_seq_len + 1,
+                                        4096,
+                                        device=device,
+                                        dtype=precision)
+
+    output1 = _causal_inference(model1, hidden_states.clone(), encoder_hidden_states.clone(), block_sizes, timesteps, precision)
+    logger.info("Finish inference for model1")
+    output2 = _causal_inference(model2, hidden_states.clone(), encoder_hidden_states.clone(), block_sizes, timesteps, precision)
+
+    # Check if outputs have the same shape
+    assert output1.shape == output2.shape, f"Output shapes don't match: {output1.shape} vs {output2.shape}"
+    assert output1.dtype == output2.dtype, f"Output dtype don't match: {output1.dtype} vs {output2.dtype}"
+
+    logger.info("Output 1 Sum: %s", output1.float().sum().item())
+    logger.info("Output 2 Sum: %s", output2.float().sum().item())
+
+    # Check if outputs are similar (allowing for small numerical differences)
+    max_diff = torch.max(torch.abs(output1 - output2))
+    mean_diff = torch.mean(torch.abs(output1 - output2))
+    logger.info("Max Diff: %s", max_diff.item())
+    logger.info("Mean Diff: %s", mean_diff.item())
+    assert max_diff < 1e-4, f"Maximum difference between outputs: {max_diff.item()}"
+    # mean diff
+    assert mean_diff < 1e-4, f"Mean difference between outputs: {mean_diff.item()}"
+
+def _causal_inference(transformer, latents, prompt_embeds, block_sizes, timesteps, target_dtype):
+    forward_batch = ForwardBatch(
+        data_type="dummy",
+    )
+    start_index = 0
+    pos_start_base = 0
+    frame_seq_length = latents.shape[-1] * latents.shape[-2] // (WanVideoConfig().arch_config.patch_size[-1] * WanVideoConfig().arch_config.patch_size[-2])
+    seq_len = frame_seq_length * latents.shape[2]
+    kv_cache1 = _initialize_kv_cache(transformer, batch_size=latents.shape[0],
+                                      kv_cache_size=frame_seq_length * latents.shape[2],
+                                      dtype=target_dtype,
+                                      device=latents.device)
+    crossattn_cache = _initialize_crossattn_cache(
+                transformer,
+                batch_size=latents.shape[0],
+                max_text_len=WanVideoConfig().arch_config.text_len,
+                dtype=target_dtype,
+                device=latents.device)
+    for current_num_frames, t_cur in zip(block_sizes, timesteps):
+        # logger.info(f"Current frame idx: {start_index}, Current timestep: {t_cur}")
+        # logger.info(f"k cache sum: {sum(kv_cache['k'].float().sum().item() for kv_cache in kv_cache1)}, v cache sum: {sum(kv_cache['v'].float().sum().item() for kv_cache in kv_cache1)}")
+        # logger.info(f"latents sum: {latents.float().sum().item()}, encoder_hidden_states sum: {prompt_embeds.float().sum().item()}")
+        current_latents = latents[:, :, start_index:start_index +
+                                    current_num_frames, :, :]
+
+        attn_metadata = None
+
+        with set_forward_context(current_timestep=0,
+                                attn_metadata=attn_metadata,
+                                forward_batch=forward_batch):
+            # Run transformer; follow DMD stage pattern
+            t_expanded_noise = t_cur * torch.ones(
+                (current_latents.shape[0], 1),
+                device=current_latents.device,
+                dtype=torch.long)
+            if isinstance(transformer, CausalWanModel):
+                pred_noise_btchw = transformer(
+                    x=current_latents,
+                    context=prompt_embeds,
+                    t=t_expanded_noise,
+                    seq_len=seq_len,
+                    kv_cache=kv_cache1,
+                    crossattn_cache=crossattn_cache,
+                    current_start=(pos_start_base + start_index) *
+                    frame_seq_length
+                )
+            elif isinstance(transformer, CausalWanTransformer3DModel):
+                pred_noise_btchw = transformer(
+                    current_latents,
+                    prompt_embeds,
+                    t_expanded_noise,
+                    kv_cache=kv_cache1,
+                    crossattn_cache=crossattn_cache,
+                    current_start=(pos_start_base + start_index) *
+                    frame_seq_length,
+                    start_frame=start_index
+                )
+
+        # Write back and advance
+        latents[:, :, start_index:start_index +
+                current_num_frames, :, :] = pred_noise_btchw.clone()
+
+        # Re-run with context timestep to update KV cache using clean context
+        context_noise = 0
+        t_context = torch.ones([latents.shape[0]],
+                                device=latents.device,
+                                dtype=torch.long) * int(context_noise)
+        context_bcthw = pred_noise_btchw.to(target_dtype)
+        with set_forward_context(current_timestep=0,
+                                attn_metadata=attn_metadata,
+                                forward_batch=forward_batch):
+            t_expanded_context = t_context.unsqueeze(1)
+            if isinstance(transformer, CausalWanModel):
+                _ = transformer(
+                    x=context_bcthw,
+                    context=prompt_embeds,
+                    t=t_expanded_context,
+                    seq_len=seq_len,
+                    kv_cache=kv_cache1,
+                    crossattn_cache=crossattn_cache,
+                    current_start=(pos_start_base + start_index) *
+                    frame_seq_length
+                )
+            elif isinstance(transformer, CausalWanTransformer3DModel):
+                _ = transformer(
+                    context_bcthw,
+                    prompt_embeds,
+                    t_expanded_context,
+                    kv_cache=kv_cache1,
+                    crossattn_cache=crossattn_cache,
+                    current_start=(pos_start_base + start_index) *
+                    frame_seq_length,
+                    start_frame=start_index
+                )
+        start_index += current_num_frames
+
+    return latents
+
+def _initialize_kv_cache(transformer, batch_size, kv_cache_size, dtype, device) -> None:
+    """
+    Initialize a Per-GPU KV cache aligned with the Wan model assumptions.
+    """
+    kv_cache1 = []
+    if isinstance(transformer, CausalWanModel):
+        num_attention_heads = transformer.num_heads
+        attention_head_dim = transformer.dim // transformer.num_heads
+    elif isinstance(transformer, CausalWanTransformer3DModel):
+        num_attention_heads = transformer.num_attention_heads
+        attention_head_dim = transformer.attention_head_dim
+
+    for _ in range(len(transformer.blocks)):
+        kv_cache1.append({
+            "k":
+            torch.zeros([
+                batch_size, kv_cache_size, num_attention_heads,
+                attention_head_dim
+            ],
+                        dtype=dtype,
+                        device=device),
+            "v":
+            torch.zeros([
+                batch_size, kv_cache_size, num_attention_heads,
+                attention_head_dim
+            ],
+                        dtype=dtype,
+                        device=device),
+            "global_end_index":
+            torch.tensor([0], dtype=torch.long, device=device),
+            "local_end_index":
+            torch.tensor([0], dtype=torch.long, device=device),
+        })
+
+    return kv_cache1
+
+def _initialize_crossattn_cache(transformer, batch_size, max_text_len, dtype,
+                                    device) -> None:
+    """
+    Initialize a Per-GPU cross-attention cache aligned with the Wan model assumptions.
+    """
+    crossattn_cache = []
+    if isinstance(transformer, CausalWanModel):
+        num_attention_heads = transformer.num_heads
+        attention_head_dim = transformer.dim // transformer.num_heads
+    elif isinstance(transformer, CausalWanTransformer3DModel):
+        num_attention_heads = transformer.num_attention_heads
+        attention_head_dim = transformer.attention_head_dim
+
+    for _ in range(len(transformer.blocks)):
+        crossattn_cache.append({
+            "k":
+            torch.zeros([
+                batch_size, max_text_len, num_attention_heads,
+                attention_head_dim
+            ],
+                        dtype=dtype,
+                        device=device),
+            "v":
+            torch.zeros([
+                batch_size, max_text_len, num_attention_heads,
+                attention_head_dim
+            ],
+                        dtype=dtype,
+                        device=device),
+            "is_init":
+            False,
+        })
+    return crossattn_cache

fastvideo/tests/transformers/test_ori_wanvideo.py

@@ -0,0 +1,133 @@
+# SPDX-License-Identifier: Apache-2.0
+import os
+import math
+
+import numpy as np
+import pytest
+import torch
+from fastvideo.wan.modules.model import WanModel
+
+from fastvideo.configs.pipelines import PipelineConfig
+from fastvideo.forward_context import set_forward_context
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.loader.component_loader import TransformerLoader
+from fastvideo.utils import maybe_download_model
+from fastvideo.configs.models.dits import WanVideoConfig
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+
+
+logger = init_logger(__name__)
+
+os.environ["MASTER_ADDR"] = "localhost"
+os.environ["MASTER_PORT"] = "29503"
+
+BASE_MODEL_PATH = "Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
+MODEL_PATH = maybe_download_model(BASE_MODEL_PATH,
+                                  local_dir=os.path.join(
+                                      'data', BASE_MODEL_PATH))
+TRANSFORMER_PATH = os.path.join(MODEL_PATH, "transformer")
+
+
+@pytest.mark.usefixtures("distributed_setup")
+def test_ori_wan_transformer():
+    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+    precision = torch.bfloat16
+    precision_str = "bf16"
+    args = FastVideoArgs(model_path=TRANSFORMER_PATH,
+                         dit_cpu_offload=True,
+                         pipeline_config=PipelineConfig(dit_config=WanVideoConfig(), dit_precision=precision_str))
+    args.device = device
+
+    loader = TransformerLoader()
+    model2 = loader.load(TRANSFORMER_PATH, args).to(dtype=precision)
+
+    model1 = WanModel.from_pretrained(
+        "/mnt/weka/home/hao.zhang/wei/Self-Forcing/wan_models/Wan2.1-T2V-1.3B", device=device,
+        torch_dtype=precision).to(device, dtype=precision).requires_grad_(False)
+
+    total_params = sum(p.numel() for p in model1.parameters())
+    # Calculate weight sum for model1 (converting to float64 to avoid overflow)
+    weight_sum_model1 = sum(
+        p.to(torch.float64).sum().item() for p in model1.parameters())
+    # Also calculate mean for more stable comparison
+    weight_mean_model1 = weight_sum_model1 / total_params
+    logger.info("Model 1 weight sum: %s", weight_sum_model1)
+    logger.info("Model 1 weight mean: %s", weight_mean_model1)
+
+    # Calculate weight sum for model2 (converting to float64 to avoid overflow)
+    total_params_model2 = sum(p.numel() for p in model2.parameters())
+    weight_sum_model2 = sum(
+        p.to(torch.float64).sum().item() for p in model2.parameters())
+    # Also calculate mean for more stable comparison
+    weight_mean_model2 = weight_sum_model2 / total_params_model2
+    logger.info("Model 2 weight sum: %s", weight_sum_model2)
+    logger.info("Model 2 weight mean: %s", weight_mean_model2)
+
+    weight_sum_diff = abs(weight_sum_model1 - weight_sum_model2)
+    logger.info("Weight sum difference: %s", weight_sum_diff)
+    weight_mean_diff = abs(weight_mean_model1 - weight_mean_model2)
+    logger.info("Weight mean difference: %s", weight_mean_diff)
+
+    # Set both models to eval mode
+    model1 = model1.eval()
+    model2 = model2.eval()
+
+    # Create identical inputs for both models
+    batch_size = 1
+    text_seq_len = 30
+    seq_len = math.ceil((160 * 90) /
+                            (2 * 2) *
+                            21)
+
+    # Video latents [B, C, T, H, W]
+    hidden_states = torch.randn(batch_size,
+                                16,
+                                21,
+                                160,
+                                90,
+                                device=device,
+                                dtype=precision)
+
+    # Text embeddings [B, L, D] (including global token)
+    encoder_hidden_states = torch.randn(batch_size,
+                                        text_seq_len + 1,
+                                        4096,
+                                        device=device,
+                                        dtype=precision)
+
+    # Timestep
+    timestep = torch.tensor([500], device=device, dtype=precision)
+
+    forward_batch = ForwardBatch(
+        data_type="dummy",
+    )
+
+    # with torch.amp.autocast('cuda', dtype=precision):
+    output1 = model1(
+        x=hidden_states,
+        context=encoder_hidden_states,
+        t=timestep,
+        seq_len=seq_len,
+    )
+    with set_forward_context(
+            current_timestep=0,
+            attn_metadata=None,
+            forward_batch=forward_batch,
+    ):
+        output2 = model2(hidden_states=hidden_states,
+                            encoder_hidden_states=encoder_hidden_states,
+                            timestep=timestep)
+
+    # Check if outputs have the same shape
+    assert output1.shape == output2.shape, f"Output shapes don't match: {output1.shape} vs {output2.shape}"
+    assert output1.dtype == output2.dtype, f"Output dtype don't match: {output1.dtype} vs {output2.dtype}"
+
+    # Check if outputs are similar (allowing for small numerical differences)
+    max_diff = torch.max(torch.abs(output1 - output2))
+    mean_diff = torch.mean(torch.abs(output1 - output2))
+    logger.info("Max Diff: %s", max_diff.item())
+    logger.info("Mean Diff: %s", mean_diff.item())
+    assert max_diff < 1e-4, f"Maximum difference between outputs: {max_diff.item()}"
+    # mean diff
+    assert mean_diff < 1e-4, f"Mean difference between outputs: {mean_diff.item()}"

fastvideo/tests/transformers/test_train_ori_causal_wanvideo.py

@@ -0,0 +1,144 @@
+# SPDX-License-Identifier: Apache-2.0
+import os
+import math
+
+import numpy as np
+import pytest
+import torch
+from fastvideo.wan.modules.causal_model import CausalWanModel
+
+from fastvideo.configs.pipelines import PipelineConfig
+from fastvideo.forward_context import set_forward_context
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.loader.component_loader import TransformerLoader
+from fastvideo.utils import maybe_download_model
+from fastvideo.configs.models.dits import WanVideoConfig
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+
+
+logger = init_logger(__name__)
+
+os.environ["MASTER_ADDR"] = "localhost"
+os.environ["MASTER_PORT"] = "29503"
+
+BASE_MODEL_PATH = "wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers"
+MODEL_PATH = maybe_download_model(BASE_MODEL_PATH,
+                                  local_dir=os.path.join(
+                                      'data', BASE_MODEL_PATH))
+TRANSFORMER_PATH = os.path.join(MODEL_PATH, "transformer")
+
+
+@pytest.mark.usefixtures("distributed_setup")
+def test_train_ori_causal_wan_transformer():
+    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
+    precision = torch.bfloat16
+    precision_str = "bf16"
+    args = FastVideoArgs(model_path=TRANSFORMER_PATH,
+                         dit_cpu_offload=True,
+                         pipeline_config=PipelineConfig(dit_config=WanVideoConfig(), dit_precision=precision_str))
+    args.device = device
+
+    loader = TransformerLoader()
+    model2 = loader.load(TRANSFORMER_PATH, args).to(dtype=precision)
+
+    model1 = CausalWanModel.from_pretrained(
+        "/mnt/weka/home/hao.zhang/wei/Self-Forcing/wan_models/Wan2.1-T2V-1.3B", device=device,
+        torch_dtype=precision).to(device, dtype=precision).requires_grad_(False)
+    causal_state_dict = torch.load("/mnt/weka/home/hao.zhang/wei/Self-Forcing/checkpoints/self_forcing_dmd.pt")["generator_ema"]
+    new_state_dict = {}
+    for k, v in causal_state_dict.items():
+        if k.startswith("model."):
+            new_state_dict[k.replace("model.", "")] = v
+    causal_state_dict = new_state_dict
+    model1.load_state_dict(causal_state_dict)
+
+    model1.num_frame_per_block = 3
+    model2.num_frame_per_block = 3
+
+    total_params = sum(p.numel() for p in model1.parameters())
+    # Calculate weight sum for model1 (converting to float64 to avoid overflow)
+    weight_sum_model1 = sum(
+        p.to(torch.float64).sum().item() for p in model1.parameters())
+    # Also calculate mean for more stable comparison
+    weight_mean_model1 = weight_sum_model1 / total_params
+    logger.info("Model 1 weight sum: %s", weight_sum_model1)
+    logger.info("Model 1 weight mean: %s", weight_mean_model1)
+
+    # Calculate weight sum for model2 (converting to float64 to avoid overflow)
+    total_params_model2 = sum(p.numel() for p in model2.parameters())
+    weight_sum_model2 = sum(
+        p.to(torch.float64).sum().item() for p in model2.parameters())
+    # Also calculate mean for more stable comparison
+    weight_mean_model2 = weight_sum_model2 / total_params_model2
+    logger.info("Model 2 weight sum: %s", weight_sum_model2)
+    logger.info("Model 2 weight mean: %s", weight_mean_model2)
+
+    weight_sum_diff = abs(weight_sum_model1 - weight_sum_model2)
+    logger.info("Weight sum difference: %s", weight_sum_diff)
+    weight_mean_diff = abs(weight_mean_model1 - weight_mean_model2)
+    logger.info("Weight mean difference: %s", weight_mean_diff)
+
+    # Set both models to eval mode
+    model1 = model1.eval()
+    model2 = model2.eval()
+
+    # Create identical inputs for both models
+    batch_size = 1
+    text_seq_len = 30
+    seq_len = math.ceil((160 * 90) /
+                            (2 * 2) *
+                            21)
+
+    # Video latents [B, C, T, H, W]
+    hidden_states = torch.randn(batch_size,
+                                16,
+                                21,
+                                160,
+                                90,
+                                device=device,
+                                dtype=precision)
+
+    # Text embeddings [B, L, D] (including global token)
+    encoder_hidden_states = torch.randn(batch_size,
+                                        text_seq_len + 1,
+                                        4096,
+                                        device=device,
+                                        dtype=precision)
+
+    # Timestep
+    timestep = torch.randint(0, 1000, (batch_size, 21), device=device, dtype=torch.long)
+    logger.info("timestep: %s", timestep)
+
+    forward_batch = ForwardBatch(
+        data_type="dummy",
+    )
+
+    # with torch.amp.autocast('cuda', dtype=precision):
+    output1 = model1(
+        x=hidden_states,
+        context=encoder_hidden_states,
+        t=timestep,
+        seq_len=seq_len,
+    )
+    with set_forward_context(
+            current_timestep=0,
+            attn_metadata=None,
+            forward_batch=forward_batch,
+    ):
+        output2 = model2(hidden_states=hidden_states,
+                            encoder_hidden_states=encoder_hidden_states,
+                            timestep=timestep)
+
+    # Check if outputs have the same shape
+    assert output1.shape == output2.shape, f"Output shapes don't match: {output1.shape} vs {output2.shape}"
+    assert output1.dtype == output2.dtype, f"Output dtype don't match: {output1.dtype} vs {output2.dtype}"
+
+    # Check if outputs are similar (allowing for small numerical differences)
+    max_diff = torch.max(torch.abs(output1 - output2))
+    mean_diff = torch.mean(torch.abs(output1 - output2))
+    logger.info("Max Diff: %s", max_diff.item())
+    logger.info("Mean Diff: %s", mean_diff.item())
+    assert max_diff < 1e-4, f"Maximum difference between outputs: {max_diff.item()}"
+    # mean diff
+    assert mean_diff < 1e-4, f"Mean difference between outputs: {mean_diff.item()}"

fastvideo/tests/workflow/test_components_parquet_saver.py

@@ -0,0 +1,68 @@
+from pathlib import Path
+
+import numpy as np
+import pyarrow as pa
+import pyarrow.parquet as pq
+import torch
+
+from fastvideo.workflow.preprocess.components import ParquetDatasetSaver
+from fastvideo.pipelines.pipeline_batch_info import PreprocessBatch
+
+
+def _simple_record_creator(batch: PreprocessBatch) -> list[dict]:
+    # batch.latents will be converted to numpy by the saver before this call
+    assert isinstance(batch.latents, np.ndarray)
+    num = len(batch.video_file_name)
+    records = []
+    for i in range(num):
+        arr = batch.latents[i]
+        records.append({
+            "id": batch.video_file_name[i],
+            "data_bytes": arr.tobytes(),
+            "data_shape": list(arr.shape),
+        })
+    return records
+
+
+def test_parquet_dataset_saver_flush_and_last(tmp_path: Path):
+    # Schema for the simple record creator
+    schema = pa.schema([
+        pa.field("id", pa.string()),
+        pa.field("data_bytes", pa.binary()),
+        pa.field("data_shape", pa.list_(pa.int64())),
+    ])
+
+    B = 5
+    # Build a minimal PreprocessBatch
+    batch = PreprocessBatch(
+        data_type="video",
+        latents=torch.randn(B, 2),
+        prompt_embeds=[torch.randn(B, 1, 1)],
+        # Attention mask should be integer dtype in real pipelines
+        prompt_attention_mask=[torch.ones(B, 1, dtype=torch.int64)],
+    )
+    batch.video_file_name = [f"vid_{i}" for i in range(B)]
+
+    saver = ParquetDatasetSaver(
+        flush_frequency=10,  # higher than B to avoid auto-flush
+        samples_per_file=3,
+        schema=schema,
+        record_creator=_simple_record_creator,
+    )
+
+    out_dir = tmp_path / "saver_out"
+    saver.save_and_write_parquet_batch(batch, str(out_dir))
+    # First flush: should write one full file (3 rows), keep 2 in buffer
+    saver.flush_tables()
+    files = sorted(out_dir.rglob("*.parquet"))
+    assert len(files) == 1
+    assert pq.read_table(str(files[0])).num_rows == 3
+
+    # Final flush: write remainder 2 rows
+    saver.flush_tables(write_remainder=True)
+    files2 = sorted(out_dir.rglob("*.parquet"))
+    assert len(files2) == 2
+    total = sum(pq.read_table(str(f)).num_rows for f in files2)
+    assert total == 5
+
+

fastvideo/training/distillation_pipeline.py

@@ -1,6 +1,7 @@
 # SPDX-License-Identifier: Apache-2.0
 import copy
 import gc
+import json
 import os
 import time
 from abc import abstractmethod
@@ -11,6 +12,7 @@ from typing import Any
 import imageio
 import numpy as np
 import torch
+import torch.distributed as dist
 import torch.nn.functional as F
 import torchvision
 from einops import rearrange
@@ -36,9 +38,11 @@ from fastvideo.training.activation_checkpoint import (
     apply_activation_checkpointing)
 from fastvideo.training.training_pipeline import TrainingPipeline
 from fastvideo.training.training_utils import (
-    clip_grad_norm_while_handling_failing_dtensor_cases, get_scheduler,
-    load_distillation_checkpoint, save_distillation_checkpoint, shift_timestep)
-from fastvideo.utils import is_vsa_available, set_random_seed
+    EMA_FSDP, clip_grad_norm_while_handling_failing_dtensor_cases,
+    get_scheduler, load_distillation_checkpoint, save_distillation_checkpoint,
+    shift_timestep)
+from fastvideo.utils import (is_vsa_available, maybe_download_model,
+                             set_random_seed, verify_model_config_and_directory)
 
 import wandb  # isort: skip
 
@@ -87,9 +91,27 @@ class DistillationPipeline(TrainingPipeline):
         self.noise_scheduler = FlowMatchEulerDiscreteScheduler(
             shift=self.timestep_shift)
 
-        # self.transformer is the generator model
-        self.real_score_transformer = self.get_module("real_score_transformer")
-        self.fake_score_transformer = self.get_module("fake_score_transformer")
+        if training_args.real_score_model_path:
+            logger.info(
+                f"Loading real score transformer from: {training_args.real_score_model_path}"
+            )
+            self.real_score_transformer = self.load_module_from_path(
+                training_args.real_score_model_path, "transformer",
+                training_args)
+        else:
+            self.real_score_transformer = self.get_module(
+                "real_score_transformer")
+
+        if training_args.fake_score_model_path:
+            logger.info(
+                f"Loading fake score transformer from: {training_args.fake_score_model_path}"
+            )
+            self.fake_score_transformer = self.load_module_from_path(
+                training_args.fake_score_model_path, "transformer",
+                training_args)
+        else:
+            self.fake_score_transformer = self.get_module(
+                "fake_score_transformer")
 
         self.real_score_transformer.requires_grad_(False)
         self.real_score_transformer.eval()
@@ -116,10 +138,13 @@ class DistillationPipeline(TrainingPipeline):
         if fake_score_lr == 0.0:
             fake_score_lr = training_args.learning_rate
 
+        betas_str = training_args.fake_score_betas
+        betas = tuple(float(x.strip()) for x in betas_str.split(","))
+
         self.fake_score_optimizer = torch.optim.AdamW(
             fake_score_params,
             lr=fake_score_lr,
-            betas=(0.9, 0.999),
+            betas=betas,
             weight_decay=training_args.weight_decay,
             eps=1e-8,
         )
@@ -147,8 +172,19 @@ class DistillationPipeline(TrainingPipeline):
             self.training_args.pipeline_config.dmd_denoising_steps,
             dtype=torch.long,
             device=get_local_torch_device())
-        logger.info("Distillation generator model to %s denoising steps",
-                    len(self.denoising_step_list))
+
+        if training_args.warp_denoising_step:  # Warp the denoising step according to the scheduler time shift
+            timesteps = torch.cat((self.noise_scheduler.timesteps.cpu(),
+                                   torch.tensor([0],
+                                                dtype=torch.float32))).cuda()
+            self.denoising_step_list = timesteps[1000 -
+                                                 self.denoising_step_list]
+            logger.info("Warping denoising_step_list")
+
+        self.denoising_step_list = self.denoising_step_list.to(
+            get_local_torch_device())
+        logger.info("Distillation generator model to %s denoising steps: %s",
+                    len(self.denoising_step_list), self.denoising_step_list)
         self.num_train_timestep = self.noise_scheduler.num_train_timesteps
 
         self.min_timestep = int(self.training_args.min_timestep_ratio *
@@ -158,6 +194,82 @@ class DistillationPipeline(TrainingPipeline):
 
         self.real_score_guidance_scale = self.training_args.real_score_guidance_scale
 
+        self.generator_ema = None
+        if (self.training_args.ema_decay
+                is not None) and (self.training_args.ema_decay > 0.0):
+            self.generator_ema = EMA_FSDP(self.transformer,
+                                          decay=self.training_args.ema_decay)
+            logger.info(
+                f"Initialized generator EMA with decay={self.training_args.ema_decay}"
+            )
+        else:
+            logger.info("Generator EMA disabled (ema_decay <= 0.0)")
+
+    def load_module_from_path(self, model_path: str, module_type: str,
+                              training_args: "TrainingArgs"):
+        """
+        Load a module from a specific path using the same loading logic as the pipeline.
+        
+        Args:
+            model_path: Path to the model
+            module_type: Type of module to load (e.g., "transformer")
+            training_args: Training arguments
+            
+        Returns:
+            The loaded module
+        """
+        logger.info(f"Loading {module_type} from custom path: {model_path}")
+        # Set flag to prevent custom weight loading for teacher/critic models
+        training_args._loading_teacher_critic_model = True
+
+        try:
+            from fastvideo.models.loader.component_loader import (
+                PipelineComponentLoader)
+
+            # Download the model if it's a Hugging Face model ID
+            local_model_path = maybe_download_model(model_path)
+            logger.info(f"Model downloaded/found at: {local_model_path}")
+            config = verify_model_config_and_directory(local_model_path)
+
+            if module_type not in config:
+                if hasattr(self, '_extra_config_module_map'
+                           ) and module_type in self._extra_config_module_map:
+                    extra_module = self._extra_config_module_map[module_type]
+                    if extra_module in config:
+                        module_type = extra_module
+                        logger.info(f"Using {extra_module} for {module_type}")
+                    else:
+                        raise ValueError(
+                            f"Module {module_type} not found in config at {local_model_path}"
+                        )
+                else:
+                    raise ValueError(
+                        f"Module {module_type} not found in config at {local_model_path}"
+                    )
+
+            module_info = config[module_type]
+            if module_info is None:
+                raise ValueError(
+                    f"Module {module_type} has null value in config at {local_model_path}"
+                )
+
+            transformers_or_diffusers, architecture = module_info
+            component_path = os.path.join(local_model_path, module_type)
+            module = PipelineComponentLoader.load_module(
+                module_name=module_type,
+                component_model_path=component_path,
+                transformers_or_diffusers=transformers_or_diffusers,
+                fastvideo_args=training_args,
+            )
+
+            logger.info(
+                f"Successfully loaded {module_type} from {component_path}")
+            return module
+        finally:
+            # Always clean up the flag
+            if hasattr(training_args, '_loading_teacher_critic_model'):
+                delattr(training_args, '_loading_teacher_critic_model')
+
     @abstractmethod
     def initialize_validation_pipeline(self, training_args: TrainingArgs):
         """Initialize validation pipeline - must be implemented by subclasses."""
@@ -174,6 +286,110 @@ class DistillationPipeline(TrainingPipeline):
 
         return training_batch
 
+    def apply_ema_to_model(self, model):
+        """Apply EMA weights to the model for validation or inference."""
+        if self.generator_ema is not None:
+            with self.generator_ema.apply_to_model(model):
+                return model
+        return model
+
+    def get_ema_model_copy(self):
+        """Get a copy of the model with EMA weights applied."""
+        if self.generator_ema is not None:
+            ema_model = copy.deepcopy(self.transformer)
+            self.generator_ema.copy_to_unwrapped(ema_model)
+            return ema_model
+        return None
+
+    def is_ema_ready(self, current_step: int = None):
+        """Check if EMA is ready for use (after ema_start_step)."""
+        if current_step is None:
+            current_step = getattr(self, 'current_trainstep', 0)
+        return (self.generator_ema is not None
+                and current_step >= self.training_args.ema_start_step)
+
+    def save_ema_weights(self, output_dir: str, step: int):
+        """Save EMA weights separately for inference purposes."""
+        if self.generator_ema is None:
+            logger.warning("Cannot save EMA weights: EMA not initialized")
+            return
+
+        if not self.is_ema_ready():
+            logger.warning(
+                "Cannot save EMA weights: EMA not ready yet (step < ema_start_step)"
+            )
+            return
+
+        try:
+            ema_model = self.get_ema_model_copy()
+            if ema_model is None:
+                logger.warning("Failed to create EMA model copy")
+                return
+
+            ema_save_dir = os.path.join(output_dir, f"ema_checkpoint-{step}")
+            os.makedirs(ema_save_dir, exist_ok=True)
+
+            # save as diffusers format
+            from safetensors.torch import save_file
+
+            from fastvideo.training.training_utils import (
+                custom_to_hf_state_dict, gather_state_dict_on_cpu_rank0)
+            cpu_state = gather_state_dict_on_cpu_rank0(ema_model, device=None)
+
+            if self.global_rank == 0:
+                weight_path = os.path.join(
+                    ema_save_dir, "diffusion_pytorch_model.safetensors")
+                diffusers_state_dict = custom_to_hf_state_dict(
+                    cpu_state, ema_model.reverse_param_names_mapping)
+                save_file(diffusers_state_dict, weight_path)
+
+                config_dict = ema_model.hf_config
+                if "dtype" in config_dict:
+                    del config_dict["dtype"]
+                config_path = os.path.join(ema_save_dir, "config.json")
+                with open(config_path, "w") as f:
+                    json.dump(config_dict, f, indent=4)
+
+                logger.info(f"EMA weights saved to {weight_path}")
+
+            del ema_model
+
+        except Exception as e:
+            logger.error(f"Failed to save EMA weights: {str(e)}")
+
+    def get_ema_stats(self):
+        """Get EMA statistics for monitoring."""
+        if self.generator_ema is None:
+            return {
+                "ema_enabled": False,
+                "ema_decay": None,
+                "ema_start_step": self.training_args.ema_start_step,
+                "ema_ready": False,
+                "ema_step": self.current_trainstep,
+            }
+
+        return {
+            "ema_enabled": True,
+            "ema_decay": self.training_args.ema_decay,
+            "ema_start_step": self.training_args.ema_start_step,
+            "ema_ready": self.is_ema_ready(),
+            "ema_step": self.current_trainstep,
+        }
+
+    def reset_ema(self):
+        """Reset EMA to current model weights."""
+        if self.generator_ema is not None:
+            logger.info("Resetting EMA to current model weights")
+            self.generator_ema.update(self.transformer)
+            # Force update to current weights by setting decay to 0 temporarily
+            original_decay = self.generator_ema.decay
+            self.generator_ema.decay = 0.0
+            self.generator_ema.update(self.transformer)
+            self.generator_ema.decay = original_decay
+            logger.info("EMA reset completed")
+        else:
+            logger.warning("Cannot reset EMA: EMA not initialized")
+
     def _build_distill_input_kwargs(
             self, noise_input: torch.Tensor, timestep: torch.Tensor,
             text_dict: dict[str, torch.Tensor] | None,
@@ -330,6 +546,7 @@ class DistillationPipeline(TrainingPipeline):
     def _dmd_forward(self, generator_pred_video: torch.Tensor,
                      training_batch: TrainingBatch) -> torch.Tensor:
         """Compute DMD (Diffusion Model Distillation) loss."""
+        original_latent = generator_pred_video
         with torch.no_grad():
             timestep = torch.randint(0,
                                      self.num_train_timestep, [1],
@@ -354,7 +571,7 @@ class DistillationPipeline(TrainingPipeline):
 
             noisy_latent = self.noise_scheduler.add_noise(
                 generator_pred_video.flatten(0, 1), noise.flatten(0, 1),
-                timestep).unflatten(0, (1, generator_pred_video.shape[1]))
+                timestep).detach().unflatten(0, (1, generator_pred_video.shape[1]))
 
             # fake_score_transformer forward
             training_batch = self._build_distill_input_kwargs(
@@ -403,24 +620,24 @@ class DistillationPipeline(TrainingPipeline):
                 pred_real_video_uncond) * self.real_score_guidance_scale
 
             grad = (faker_score_pred_video - real_score_pred_video) / torch.abs(
-                generator_pred_video - real_score_pred_video).mean()
+                original_latent - real_score_pred_video).mean()
             grad = torch.nan_to_num(grad)
 
         dmd_loss = 0.5 * F.mse_loss(
-            generator_pred_video.float(),
-            (generator_pred_video.float() - grad.float()).detach())
+            original_latent.float(),
+            (original_latent.float() - grad.float()).detach())
 
         training_batch.dmd_latent_vis_dict.update({
             "training_batch_dmd_fwd_clean_latent":
             training_batch.latents,
             "generator_pred_video":
-            generator_pred_video,
+            original_latent.detach(),
             "real_score_pred_video":
-            real_score_pred_video,
+            real_score_pred_video.detach(),
             "faker_score_pred_video":
-            faker_score_pred_video,
+            faker_score_pred_video.detach(),
             "dmd_timestep":
-            timestep,
+            timestep.detach(),
         })
 
         return dmd_loss
@@ -517,12 +734,12 @@ class DistillationPipeline(TrainingPipeline):
             "encoder_hidden_states": self.negative_prompt_embeds,
             "encoder_attention_mask": self.negative_prompt_attention_mask,
         }
+        training_batch.unconditional_dict = unconditional_dict
 
         training_batch.dmd_latent_vis_dict = {}
         training_batch.fake_score_latent_vis_dict = {}
 
         training_batch.conditional_dict = conditional_dict
-        training_batch.unconditional_dict = unconditional_dict
         training_batch.raw_latent_shape = training_batch.latents.shape
         training_batch.latents = training_batch.latents.permute(0, 2, 1, 3, 4)
         self.video_latent_shape = training_batch.latents.shape
@@ -585,8 +802,15 @@ class DistillationPipeline(TrainingPipeline):
                     (dmd_loss / gradient_accumulation_steps).backward()
                 total_dmd_loss += dmd_loss.detach().item()
             self._clip_model_grad_norm_(batch_gen, self.transformer)
+            for param in self.transformer.parameters():
+                # check if the gradient is not None and not zero
+                assert param.grad is not None and param.grad.abs().sum() > 0
             self.optimizer.step()
             self.optimizer.zero_grad(set_to_none=True)
+
+            if self.generator_ema is not None:
+                self.generator_ema.update(self.transformer)
+
             avg_dmd_loss = torch.tensor(total_dmd_loss /
                                         gradient_accumulation_steps,
                                         device=self.device)
@@ -610,6 +834,9 @@ class DistillationPipeline(TrainingPipeline):
             fake_score_latent_vis_dict.update(
                 batch_fake.fake_score_latent_vis_dict)
         self._clip_model_grad_norm_(batch_fake, self.fake_score_transformer)
+        for param in self.fake_score_transformer.parameters():
+            # check if the gradient is not None and not zero
+            assert param.grad is not None and param.grad.abs().sum() > 0
         self.fake_score_optimizer.step()
         self.fake_score_lr_scheduler.step()
         self.lr_scheduler.step()
@@ -637,7 +864,8 @@ class DistillationPipeline(TrainingPipeline):
             self.transformer, self.fake_score_transformer, self.global_rank,
             self.training_args.resume_from_checkpoint, self.optimizer,
             self.fake_score_optimizer, self.train_dataloader, self.lr_scheduler,
-            self.fake_score_lr_scheduler, self.noise_random_generator)
+            self.fake_score_lr_scheduler, self.noise_random_generator,
+            self.generator_ema)
 
         if resumed_step > 0:
             self.init_steps = resumed_step
@@ -668,6 +896,14 @@ class DistillationPipeline(TrainingPipeline):
             sum(p.numel()
                 for p in self.fake_score_transformer.parameters()) / 1e9)
 
+        if self.generator_ema is not None:
+            logger.info("  Generator EMA enabled with decay: %s",
+                        self.training_args.ema_decay)
+            logger.info("  Generator EMA start step: %s",
+                        self.training_args.ema_start_step)
+        else:
+            logger.info("  Generator EMA disabled")
+
     @torch.no_grad()
     def _log_validation(self, transformer, training_args, global_step) -> None:
         training_args.inference_mode = True
@@ -699,6 +935,18 @@ class DistillationPipeline(TrainingPipeline):
 
         transformer.eval()
 
+        # Optionally use EMA model for validation if available and ready
+        use_ema_for_validation = (self.training_args.use_ema
+                                  and self.is_ema_ready(global_step))
+        if use_ema_for_validation:
+            logger.info("Using EMA model for validation")
+            validation_transformer = self.transformer
+            ema_context = self.generator_ema.apply_to_model(
+                validation_transformer)
+        else:
+            validation_transformer = transformer
+            ema_context = None
+
         validation_steps = training_args.validation_sampling_steps.split(",")
         validation_steps = [int(step) for step in validation_steps]
         validation_steps = [step for step in validation_steps if step > 0]
@@ -714,50 +962,98 @@ class DistillationPipeline(TrainingPipeline):
             step_videos: list[np.ndarray] = []
             step_captions: list[str] = []
 
-            for validation_batch in validation_dataloader:
-                batch = self._prepare_validation_batch(sampling_param,
-                                                       training_args,
-                                                       validation_batch,
-                                                       num_inference_steps)
+            if ema_context is not None:
+                with ema_context:
+                    for validation_batch in validation_dataloader:
+                        batch = self._prepare_validation_batch(
+                            sampling_param, training_args, validation_batch,
+                            num_inference_steps)
 
-                negative_prompt = batch.negative_prompt
-                batch_negative = ForwardBatch(
-                    data_type="video",
-                    prompt=negative_prompt,
-                    prompt_embeds=[],
-                    prompt_attention_mask=[],
-                )
-                result_batch = self.validation_pipeline.prompt_encoding_stage(  # type: ignore
-                    batch_negative, training_args)
-                self.negative_prompt_embeds, self.negative_prompt_attention_mask = result_batch.prompt_embeds[
-                    0], result_batch.prompt_attention_mask[0]
+                        negative_prompt = batch.negative_prompt
+                        batch_negative = ForwardBatch(
+                            data_type="video",
+                            prompt=negative_prompt,
+                            prompt_embeds=[],
+                            prompt_attention_mask=[],
+                        )
+                        result_batch = self.validation_pipeline.prompt_encoding_stage(  # type: ignore
+                            batch_negative, training_args)
+                        self.negative_prompt_embeds, self.negative_prompt_attention_mask = result_batch.prompt_embeds[
+                            0], result_batch.prompt_attention_mask[0]
 
-                logger.info("rank: %s: rank_in_sp_group: %s, batch.prompt: %s",
+                        logger.info(
+                            "rank: %s: rank_in_sp_group: %s, batch.prompt: %s",
                             self.global_rank,
                             self.rank_in_sp_group,
                             batch.prompt,
                             local_main_process_only=False)
 
-                assert batch.prompt is not None and isinstance(
-                    batch.prompt, str)
-                step_captions.append(batch.prompt)
+                        assert batch.prompt is not None and isinstance(
+                            batch.prompt, str)
+                        step_captions.append(batch.prompt)
 
-                # Run validation inference
-                with torch.no_grad():
-                    output_batch = self.validation_pipeline.forward(
-                        batch, training_args)
-                samples = output_batch.output
-                if self.rank_in_sp_group != 0:
-                    continue
+                        # Run validation inference
+                        with torch.no_grad():
+                            output_batch = self.validation_pipeline.forward(
+                                batch, training_args)
+                        samples = output_batch.output
+                        if self.rank_in_sp_group != 0:
+                            continue
 
-                # Process outputs
-                video = rearrange(samples, "b c t h w -> t b c h w")
-                frames = []
-                for x in video:
-                    x = torchvision.utils.make_grid(x, nrow=6)
-                    x = x.transpose(0, 1).transpose(1, 2).squeeze(-1)
-                    frames.append((x * 255).numpy().astype(np.uint8))
-                step_videos.append(frames)
+                        # Process outputs
+                        video = rearrange(samples, "b c t h w -> t b c h w")
+                        frames = []
+                        for x in video:
+                            x = torchvision.utils.make_grid(x, nrow=6)
+                            x = x.transpose(0, 1).transpose(1, 2).squeeze(-1)
+                            frames.append((x * 255).numpy().astype(np.uint8))
+                        step_videos.append(frames)
+            else:
+                # Use original transformer without EMA
+                for validation_batch in validation_dataloader:
+                    batch = self._prepare_validation_batch(
+                        sampling_param, training_args, validation_batch,
+                        num_inference_steps)
+
+                    negative_prompt = batch.negative_prompt
+                    batch_negative = ForwardBatch(
+                        data_type="video",
+                        prompt=negative_prompt,
+                        prompt_embeds=[],
+                        prompt_attention_mask=[],
+                    )
+                    result_batch = self.validation_pipeline.prompt_encoding_stage(  # type: ignore
+                        batch_negative, training_args)
+                    self.negative_prompt_embeds, self.negative_prompt_attention_mask = result_batch.prompt_embeds[
+                        0], result_batch.prompt_attention_mask[0]
+
+                    logger.info(
+                        "rank: %s: rank_in_sp_group: %s, batch.prompt: %s",
+                        self.global_rank,
+                        self.rank_in_sp_group,
+                        batch.prompt,
+                        local_main_process_only=False)
+
+                    assert batch.prompt is not None and isinstance(
+                        batch.prompt, str)
+                    step_captions.append(batch.prompt)
+
+                    # Run validation inference
+                    with torch.no_grad():
+                        output_batch = self.validation_pipeline.forward(
+                            batch, training_args)
+                    samples = output_batch.output
+                    if self.rank_in_sp_group != 0:
+                        continue
+
+                    # Process outputs
+                    video = rearrange(samples, "b c t h w -> t b c h w")
+                    frames = []
+                    for x in video:
+                        x = torchvision.utils.make_grid(x, nrow=6)
+                        x = x.transpose(0, 1).transpose(1, 2).squeeze(-1)
+                        frames.append((x * 255).numpy().astype(np.uint8))
+                    step_videos.append(frames)
 
             # Log validation results for this step
             world_group = get_world_group()
@@ -834,16 +1130,16 @@ class DistillationPipeline(TrainingPipeline):
                                                         latents.dtype)
                 else:
                     latents += self.vae.shift_factor
-            with torch.autocast("cuda", dtype=torch.bfloat16):
-                video = self.vae.decode(latents)
-            video = (video / 2 + 0.5).clamp(0, 1)
-            video = video.cpu().float()
-            video = video.permute(0, 2, 1, 3, 4)
-            video = (video * 255).numpy().astype(np.uint8)
-            wandb_loss_dict[latent_key] = wandb.Video(
-                video, fps=24, format="mp4")  # change to 16 for Wan2.1
-            # Clean up references
-            del video, latents
+                with torch.autocast("cuda", dtype=torch.bfloat16):
+                    video = self.vae.decode(latents)
+                video = (video / 2 + 0.5).clamp(0, 1)
+                video = video.cpu().float()
+                video = video.permute(0, 2, 1, 3, 4)
+                video = (video * 255).numpy().astype(np.uint8)
+                wandb_loss_dict[latent_key] = wandb.Video(
+                    video, fps=24, format="mp4")  # change to 16 for Wan2.1
+                # Clean up references
+                del video, latents
 
         # Process DMD training data if available - use decode_stage instead of self.vae.decode
         if 'generator_pred_video' in dmd_latents_vis_dict:
@@ -904,6 +1200,10 @@ class DistillationPipeline(TrainingPipeline):
             device="cpu").manual_seed(self.seed)
         logger.info("Initialized random seeds with seed: %s", seed)
 
+        # Initialize current_trainstep for EMA ready checks
+        #TODO: check if needed
+        self.current_trainstep = self.init_steps
+
         # Resume from checkpoint if specified (this will restore random states)
         if self.training_args.resume_from_checkpoint:
             self._resume_from_checkpoint()
@@ -947,6 +1247,14 @@ class DistillationPipeline(TrainingPipeline):
             self.current_trainstep = step
             training_batch.current_vsa_sparsity = current_vsa_sparsity
 
+            if (step >= self.training_args.ema_start_step) and \
+                    (self.generator_ema is None) and (self.training_args.ema_decay > 0):
+                self.generator_ema = EMA_FSDP(
+                    self.transformer, decay=self.training_args.ema_decay)
+                logger.info(
+                    f"Created generator EMA at step {step} with decay={self.training_args.ema_decay}"
+                )
+
             with torch.autocast("cuda", dtype=torch.bfloat16):
                 training_batch = self.train_one_step(training_batch)
 
@@ -960,11 +1268,19 @@ class DistillationPipeline(TrainingPipeline):
             avg_step_time = sum(step_times) / len(step_times)
 
             progress_bar.set_postfix({
-                "total_loss": f"{total_loss:.4f}",
-                "generator_loss": f"{generator_loss:.4f}",
-                "fake_score_loss": f"{fake_score_loss:.4f}",
-                "step_time": f"{step_time:.2f}s",
-                "grad_norm": grad_norm,
+                "total_loss":
+                f"{total_loss:.4f}",
+                "generator_loss":
+                f"{generator_loss:.4f}",
+                "fake_score_loss":
+                f"{fake_score_loss:.4f}",
+                "step_time":
+                f"{step_time:.2f}s",
+                "grad_norm":
+                grad_norm,
+                "ema":
+                "✓" if (self.generator_ema is not None and self.is_ema_ready())
+                else "✗",
             })
             progress_bar.update(1)
 
@@ -992,6 +1308,15 @@ class DistillationPipeline(TrainingPipeline):
                 if use_vsa:
                     log_data["VSA_train_sparsity"] = current_vsa_sparsity
 
+                if self.generator_ema is not None:
+                    log_data["ema_enabled"] = True
+                    log_data["ema_decay"] = self.training_args.ema_decay
+                else:
+                    log_data["ema_enabled"] = False
+
+                ema_stats = self.get_ema_stats()
+                log_data.update(ema_stats)
+
                 if training_batch.dmd_latent_vis_dict:
                     dmd_additional_logs = {
                         "generator_timestep":
@@ -1023,7 +1348,8 @@ class DistillationPipeline(TrainingPipeline):
                     self.global_rank, self.training_args.output_dir, step,
                     self.optimizer, self.fake_score_optimizer,
                     self.train_dataloader, self.lr_scheduler,
-                    self.fake_score_lr_scheduler, self.noise_random_generator)
+                    self.fake_score_lr_scheduler, self.noise_random_generator,
+                    self.generator_ema)
 
                 if self.transformer:
                     self.transformer.train()
@@ -1040,7 +1366,11 @@ class DistillationPipeline(TrainingPipeline):
                                              self.global_rank,
                                              self.training_args.output_dir,
                                              f"{step}_weight_only",
-                                             only_save_generator_weight=True)
+                                             only_save_generator_weight=True,
+                                             generator_ema=self.generator_ema)
+
+                if self.training_args.use_ema and self.is_ema_ready():
+                    self.save_ema_weights(self.training_args.output_dir, step)
 
             if self.training_args.log_validation and step % self.training_args.validation_steps == 0:
                 if self.training_args.log_visualization:
@@ -1060,7 +1390,11 @@ class DistillationPipeline(TrainingPipeline):
             self.training_args.output_dir, self.training_args.max_train_steps,
             self.optimizer, self.fake_score_optimizer, self.train_dataloader,
             self.lr_scheduler, self.fake_score_lr_scheduler,
-            self.noise_random_generator)
+            self.noise_random_generator, self.generator_ema)
+
+        if self.training_args.use_ema and self.is_ema_ready():
+            self.save_ema_weights(self.training_args.output_dir,
+                                  self.training_args.max_train_steps)
 
         if get_sp_group():
-            cleanup_dist_env_and_memory()
+            cleanup_dist_env_and_memory()

fastvideo/training/ode_causal_pipeline.py

@@ -3,19 +3,22 @@ import sys
 from copy import deepcopy
 from typing import cast
 
+
 import torch
 import torch.nn.functional as F
-
-from fastvideo.dataset.dataloader.schema import pyarrow_schema_ode_trajectory
+import numpy as np
+import wandb
+from fastvideo.dataset.dataloader.schema import pyarrow_schema_ode_trajectory_text_only
 from fastvideo.distributed import get_local_torch_device
 from fastvideo.fastvideo_args import FastVideoArgs, TrainingArgs
 from fastvideo.forward_context import set_forward_context
 from fastvideo.logger import init_logger
-from fastvideo.models.schedulers.scheduling_flow_match_euler_discrete import (
-    FlowMatchEulerDiscreteScheduler)
+from fastvideo.models.schedulers.scheduling_self_forcing_flow_match import (
+    SelfForcingFlowMatchScheduler)
 from fastvideo.pipelines.basic.wan.wan_causal_dmd_pipeline import (
     WanCausalDMDPipeline)
 from fastvideo.training.training_pipeline import TrainingPipeline
+from fastvideo.pipelines.pipeline_batch_info import TrainingBatch
 from fastvideo.training.training_utils import (
     clip_grad_norm_while_handling_failing_dtensor_cases)
 
@@ -36,11 +39,15 @@ class ODEInitTrainingPipeline(TrainingPipeline):
 
     def initialize_pipeline(self, fastvideo_args: FastVideoArgs):
         # Match the preprocess/generation scheduler for consistent stepping
-        self.modules["scheduler"] = FlowMatchEulerDiscreteScheduler(
-            shift=fastvideo_args.pipeline_config.flow_shift)
+        self.modules["scheduler"] = SelfForcingFlowMatchScheduler(
+            shift=fastvideo_args.pipeline_config.flow_shift,
+            sigma_min=0.0,
+            extra_one_step=True)
+        self.modules["scheduler"].set_timesteps(num_inference_steps=1000,
+                                                training=True)
 
     def set_schemas(self):
-        self.train_dataset_schema = pyarrow_schema_ode_trajectory
+        self.train_dataset_schema = pyarrow_schema_ode_trajectory_text_only
 
     def initialize_training_pipeline(self, training_args: TrainingArgs):
         super().initialize_training_pipeline(training_args)
@@ -51,14 +58,16 @@ class ODEInitTrainingPipeline(TrainingPipeline):
 
         self.timestep_shift = self.training_args.pipeline_config.flow_shift
         assert self.timestep_shift == 5.0, "flow_shift must be 5.0"
-        self.noise_scheduler = FlowMatchEulerDiscreteScheduler(
-            shift=self.timestep_shift)
+        self.noise_scheduler = SelfForcingFlowMatchScheduler(
+            shift=self.timestep_shift, sigma_min=0.0, extra_one_step=True)
+        self.noise_scheduler.set_timesteps(num_inference_steps=1000,
+                                           training=True)
 
         # logger.info(f"ARG dmd_denoising_steps: {training_args.pipeline_config.dmd_denoising_steps}")
         logger.info(
             f"ARG dmd_denoising_steps: {self.training_args.pipeline_config.dmd_denoising_steps}"
         )
-        self.dmd_denoising_steps = torch.tensor([1000, 750, 500, 0],
+        self.dmd_denoising_steps = torch.tensor([1000, 750, 500, 250],
                                                 dtype=torch.long,
                                                 device=get_local_torch_device())
         # self.dmd_denoising_steps = torch.tensor([1000, 750, 500, 250], dtype=torch.long, device=get_local_torch_device())
@@ -71,9 +80,9 @@ class ODEInitTrainingPipeline(TrainingPipeline):
                                                  self.dmd_denoising_steps]
             logger.info(
                 f"warped self.dmd_denoising_steps: {self.dmd_denoising_steps}")
-            assert False, "warp_denoising_step must be false"
+            # assert False, "warp_denoising_step must be false"
         else:
-            # assert False, "warp_denoising_step must be true"
+            assert False, "warp_denoising_step must be true"
             logger.info("not warped")
         self.dmd_denoising_steps = self.dmd_denoising_steps.to(
             get_local_torch_device())
@@ -98,7 +107,8 @@ class ODEInitTrainingPipeline(TrainingPipeline):
         args_copy.inference_mode = True
         # Warm start validation with current transformer
         self.validation_pipeline = WanCausalDMDPipeline.from_pretrained(
-            training_args.model_path,
+            # training_args.model_path,
+            "wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers",
             args=args_copy,  # type: ignore
             inference_mode=True,
             loaded_modules={
@@ -191,7 +201,8 @@ class ODEInitTrainingPipeline(TrainingPipeline):
         self, traj_latents: torch.Tensor, traj_timesteps: torch.Tensor,
         encoder_hidden_states: torch.Tensor,
         encoder_attention_mask: torch.Tensor
-    ) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+    ) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor, dict[str, torch.Tensor]]:
+        latent_vis_dict = {}
         device = get_local_torch_device()
         target_latent = traj_latents[:, -1]
 
@@ -211,8 +222,12 @@ class ODEInitTrainingPipeline(TrainingPipeline):
             # self._cached_closest_idx_per_dmd = distances_ks.argmin(dim=1).to(torch.long).cpu()  # [K]
             self._cached_closest_idx_per_dmd = torch.tensor(
                 [0, 12, 24, 36], dtype=torch.long).cpu()
-            # logger.info(f"self._cached_closest_idx_per_dmd: {self._cached_closest_idx_per_dmd}")
-            # logger.info(f"corresponding timesteps: {self.noise_scheduler.timesteps[self._cached_closest_idx_per_dmd]}")
+            logger.info(
+                f"self._cached_closest_idx_per_dmd: {self._cached_closest_idx_per_dmd}"
+            )
+            logger.info(
+                f"corresponding timesteps: {self.noise_scheduler.timesteps[self._cached_closest_idx_per_dmd]}"
+            )
 
         # logger.info(f"traj_latents: {traj_latents.shape}")
         # Select the K indexes from traj_latents using self._cached_closest_idx_per_dmd
@@ -222,7 +237,6 @@ class ODEInitTrainingPipeline(TrainingPipeline):
             traj_latents,
             dim=1,
             index=self._cached_closest_idx_per_dmd.to(traj_latents.device))
-        # logger.info(f"relevant_traj_latents: {relevant_traj_latents.shape}")
         # assert relevant_traj_latents.shape[0] == 1
 
         indexes = self._get_timestep(  # [B, num_frames]
@@ -232,8 +246,9 @@ class ODEInitTrainingPipeline(TrainingPipeline):
             num_frames,
             3,
             uniform_timestep=False)
+        logger.info(f"indexes: {indexes.shape}")
+        logger.info(f"indexes: {indexes}")
         # noisy_input = relevant_traj_latents[indexes]
-        # logger.info(f"indexes: {indexes.shape}")
         noisy_input = torch.gather(
             relevant_traj_latents,
             dim=1,
@@ -264,10 +279,14 @@ class ODEInitTrainingPipeline(TrainingPipeline):
         # logger.info(f"timestep: {timestep}")
 
         # Prepare inputs for transformer
+        latent_vis_dict["noisy_input"] = noisy_input.permute(0, 2, 1, 3, 4).detach().clone().cpu()
+        latent_vis_dict["x0"] = target_latent.permute(0, 2, 1, 3, 4).detach().clone().cpu()
+
+        model_dtype = next(self.transformer.parameters()).dtype
         input_kwargs = {
             "hidden_states": noisy_input.permute(0, 2, 1, 3, 4),
             "encoder_hidden_states": encoder_hidden_states,
-            "timestep": timestep.to(device, dtype=torch.bfloat16),
+            "timestep": timestep.to(device, dtype=model_dtype),
             "encoder_attention_mask": encoder_attention_mask,
             "return_dict": False,
         }
@@ -281,17 +300,18 @@ class ODEInitTrainingPipeline(TrainingPipeline):
             noise_pred = noise_pred[0]
 
         from fastvideo.models.utils import pred_noise_to_pred_video
-        noise_pred = pred_noise_to_pred_video(
+        pred_video = pred_noise_to_pred_video(
             pred_noise=noise_pred.flatten(0, 1),
             noise_input_latent=noisy_input.flatten(0, 1),
-            timestep=timestep.flatten(0, 1),
+            timestep=timestep.to(dtype=model_dtype).flatten(0, 1),
             scheduler=self.modules["scheduler"]).unflatten(
                 0, noise_pred.shape[:2])
+        latent_vis_dict["pred_video"] = pred_video.permute(0, 2, 1, 3, 4).detach().clone().cpu()
 
         # noisy_input = pred_noise_to_pred_video(noise_pred, noisy_input, t, self.modules["scheduler"])
         # next_latent_pred = self.modules["scheduler"].step(
         #     noise_pred, t, current_latents, return_dict=False)[0]
-        return noise_pred, target_latent, timestep
+        return pred_video, target_latent, timestep, latent_vis_dict
 
     def train_one_step(self, training_batch):  # type: ignore[override]
         self.transformer.train()
@@ -330,8 +350,10 @@ class ODEInitTrainingPipeline(TrainingPipeline):
             #     t = t.long()
 
             # Forward to predict next latent by stepping scheduler with predicted noise
-            noise_pred, target_latent, t = self._step_predict_next_latent(
+            noise_pred, target_latent, t, latent_vis_dict = self._step_predict_next_latent(
                 traj_latents, traj_timesteps, text_embeds, text_attention_mask)
+            
+            training_batch.latent_vis_dict.update(latent_vis_dict)
 
             mask = t != 0
 
@@ -369,6 +391,35 @@ class ODEInitTrainingPipeline(TrainingPipeline):
         training_batch.grad_norm = grad_value
         return training_batch
 
+    def visualize_intermediate_latents(self, training_batch: TrainingBatch,
+                                       training_args: TrainingArgs, step: int):
+        """Add visualization data to wandb logging and save frames to disk."""
+        wandb_loss_dict = {}
+        latents_vis_dict = training_batch.latent_vis_dict
+        latent_log_keys = ['noisy_input', 'x0', 'pred_video']
+        for latent_key in latent_log_keys:
+            assert latent_key in latents_vis_dict and latents_vis_dict[latent_key] is not None
+            latent = latents_vis_dict[latent_key]
+            pixel_latent = self.validation_pipeline.decoding_stage.decode(latent, training_args)
+
+            video = pixel_latent.cpu().float()
+            video = video.permute(0, 2, 1, 3, 4)
+            video = (video * 255).numpy().astype(np.uint8)
+            wandb_loss_dict[latent_key] = wandb.Video(
+                video, fps=16, format="mp4")  # change to 16 for Wan2.1
+            # Clean up references
+            del video, pixel_latent, latent
+
+        # Log to wandb
+        if self.global_rank == 0:
+            wandb.log(wandb_loss_dict, step=step)
+
+
+        # dmd_latents_vis_dict = training_batch.dmd_latent_vis_dict
+        # fake_score_latents_vis_dict = training_batch.fake_score_latent_vis_dict
+        # fake_score_log_keys = ['generator_pred_video']
+        # dmd_log_keys = ['faker_score_pred_video', 'real_score_pred_video']
+
 
 def main(args) -> None:
     logger.info("Starting ODE-init training pipeline...")

fastvideo/training/training_pipeline.py

@@ -22,6 +22,7 @@ from tqdm.auto import tqdm
 import fastvideo.envs as envs
 from fastvideo.attention.backends.video_sparse_attn import (
     VideoSparseAttentionMetadataBuilder)
+# from fastvideo.attention.backends.vmoba import VideoMobaAttentionMetadataBuilder
 from fastvideo.configs.sample import SamplingParam
 from fastvideo.dataset import build_parquet_map_style_dataloader
 from fastvideo.dataset.dataloader.schema import pyarrow_schema_t2v
@@ -41,11 +42,14 @@ from fastvideo.training.training_utils import (
     compute_density_for_timestep_sampling, get_scheduler, get_sigmas,
     load_checkpoint, normalize_dit_input, save_checkpoint,
     shard_latents_across_sp)
+# from fastvideo.utils import (is_vmoba_available, is_vsa_available,
+#                              set_random_seed, shallow_asdict)
 from fastvideo.utils import is_vsa_available, set_random_seed, shallow_asdict
 
 import wandb  # isort: skip
 
 vsa_available = is_vsa_available()
+# vmoba_available = is_vmoba_available()
 
 logger = init_logger(__name__)
 
@@ -117,10 +121,14 @@ class TrainingPipeline(LoRAPipeline, ABC):
         params_to_optimize = self.transformer.parameters()
         params_to_optimize = list(
             filter(lambda p: p.requires_grad, params_to_optimize))
+        # Parse betas from string format "beta1,beta2"
+        betas_str = training_args.betas
+        betas = tuple(float(x.strip()) for x in betas_str.split(","))
+
         self.optimizer = torch.optim.AdamW(
             params_to_optimize,
             lr=training_args.learning_rate,
-            betas=(0.9, 0.999),
+            betas=betas,
             weight_decay=training_args.weight_decay,
             eps=1e-8,
         )
@@ -272,6 +280,20 @@ class TrainingPipeline(LoRAPipeline, ABC):
                 patch_size=patch_size,
                 VSA_sparsity=current_vsa_sparsity,
                 device=get_local_torch_device())
+        # elif vmoba_available and envs.FASTVIDEO_ATTENTION_BACKEND == "VMOBA_ATTN":
+        #     moba_params = self.training_args.moba_config.copy()
+        #     moba_params.update({
+        #         "current_timestep":
+        #         training_batch.timesteps,
+        #         "raw_latent_shape":
+        #         training_batch.raw_latent_shape[2:5],
+        #         "patch_size":
+        #         self.training_args.pipeline_config.dit_config.patch_size,
+        #         "device":
+        #         get_local_torch_device(),
+        #     })
+        #     training_batch.attn_metadata = VideoMobaAttentionMetadataBuilder(
+        #     ).build(**moba_params)
         else:
             training_batch.attn_metadata = None
 
@@ -296,6 +318,7 @@ class TrainingPipeline(LoRAPipeline, ABC):
 
     def _transformer_forward_and_compute_loss(
             self, training_batch: TrainingBatch) -> TrainingBatch:
+        # if vsa_available and envs.FASTVIDEO_ATTENTION_BACKEND == "VIDEO_SPARSE_ATTN" or vmoba_available and envs.FASTVIDEO_ATTENTION_BACKEND == "VMOBA_ATTN":
         if vsa_available and envs.FASTVIDEO_ATTENTION_BACKEND == "VIDEO_SPARSE_ATTN":
             assert training_batch.attn_metadata is not None
         else:
@@ -462,6 +485,9 @@ class TrainingPipeline(LoRAPipeline, ABC):
                 current_decay_times = min(step // vsa_decay_interval_steps,
                                           vsa_sparsity // vsa_decay_rate)
                 current_vsa_sparsity = current_decay_times * vsa_decay_rate
+            # elif vmoba_available:
+            #     # TODO: add vmoba sparsity scheduling here
+            #     current_vsa_sparsity = 0.0
             else:
                 current_vsa_sparsity = 0.0
 
@@ -503,6 +529,10 @@ class TrainingPipeline(LoRAPipeline, ABC):
                 self.transformer.train()
                 self.sp_group.barrier()
             if self.training_args.log_validation and step % self.training_args.validation_steps == 0:
+                if self.training_args.log_visualization:
+                    self.visualize_intermediate_latents(training_batch,
+                                                        self.training_args,
+                                                        step)
                 self._log_validation(self.transformer, self.training_args, step)
                 gpu_memory_usage = torch.cuda.memory_allocated() / 1024**2
                 trainable_params = round(
@@ -704,3 +734,10 @@ class TrainingPipeline(LoRAPipeline, ABC):
         # Re-enable gradients for training
         training_args.inference_mode = False
         transformer.train()
+
+    def visualize_intermediate_latents(self, training_batch: TrainingBatch,
+                                       training_args: TrainingArgs, step: int):
+        """Add visualization data to wandb logging and save frames to disk."""
+        raise NotImplementedError(
+            "Visualize intermediate latents is not implemented for training pipeline"
+        )

fastvideo/training/training_utils.py

@@ -202,6 +202,7 @@ def save_distillation_checkpoint(generator_transformer,
                                  generator_scheduler=None,
                                  fake_score_scheduler=None,
                                  noise_generator=None,
+                                 generator_ema=None,
                                  only_save_generator_weight=False) -> None:
     """
     Save distillation checkpoint with both generator and fake_score models.
@@ -233,6 +234,8 @@ def save_distillation_checkpoint(generator_transformer,
         if generator_scheduler is not None:
             generator_states["scheduler"] = SchedulerWrapper(
                 generator_scheduler)
+        if generator_ema is not None:
+            generator_states["ema"] = generator_ema.state_dict()
 
         generator_dcp_dir = os.path.join(save_dir, "distributed_checkpoint",
                                          "generator")
@@ -406,7 +409,8 @@ def load_distillation_checkpoint(generator_transformer,
                                  dataloader=None,
                                  generator_scheduler=None,
                                  fake_score_scheduler=None,
-                                 noise_generator=None) -> int:
+                                 noise_generator=None,
+                                 generator_ema=None) -> int:
     """
     Load distillation checkpoint with both generator and fake_score models.
     Returns the step number from which training should resume.
@@ -460,6 +464,20 @@ def load_distillation_checkpoint(generator_transformer,
         end_time - begin_time,
         local_main_process_only=False)
 
+    # Load EMA state if available and generator_ema is provided
+    if generator_ema is not None:
+        try:
+            ema_state = generator_states.get("ema")
+            if ema_state is not None:
+                generator_ema.load_state_dict(ema_state)
+                logger.info("rank: %s, generator EMA state loaded successfully",
+                            rank)
+            else:
+                logger.info("rank: %s, no EMA state found in checkpoint", rank)
+        except Exception as e:
+            logger.warning("rank: %s, failed to load EMA state: %s", rank,
+                           str(e))
+
     # Load critic distributed checkpoint
     critic_dcp_dir = os.path.join(checkpoint_path, "distributed_checkpoint",
                                   "critic")
@@ -1282,3 +1300,169 @@ def get_scheduler(
                          num_warmup_steps=num_warmup_steps,
                          num_training_steps=num_training_steps,
                          last_epoch=last_epoch)
+
+
+class EMA_FSDP:
+    """
+    FSDP2-friendly EMA with two modes:
+      - mode="local_shard" (default): maintain float32 CPU EMA of local parameter shards on every rank.
+        Provides a context manager to temporarily swap EMA weights into the live model for teacher forward.
+      - mode="rank0_full": maintain a consolidated float32 CPU EMA of full parameters on rank 0 only
+        using gather_state_dict_on_cpu_rank0(). Useful for checkpoint export; not for teacher forward.
+
+    Usage (local_shard for CM teacher):
+      ema = EMA_FSDP(model, decay=0.999, mode="local_shard")
+      for step in ...:
+          ema.update(model)
+      with ema.apply_to_model(model):
+          with torch.no_grad():
+              y_teacher = model(...)
+
+    Usage (rank0_full for export):
+      ema = EMA_FSDP(model, decay=0.999, mode="rank0_full")
+      ema.update(model)
+      ema.state_dict()  # on rank 0
+    """
+
+    def __init__(self, module, decay: float = 0.999, mode: str = "local_shard"):
+        self.decay = float(decay)
+        self.mode = mode
+        self.shadow: dict[str, torch.Tensor] = {}
+        self.rank = dist.get_rank() if dist.is_initialized() else 0
+        if self.mode not in {"local_shard", "rank0_full"}:
+            raise ValueError(f"Unsupported EMA_FSDP mode: {self.mode}")
+        self._init_shadow(module)
+
+    @staticmethod
+    def _to_local_tensor(t: torch.Tensor) -> torch.Tensor:
+        # DTensor-aware to_local fetch; fall back to raw tensor
+        try:
+            from torch.distributed.tensor import DTensor  # type: ignore
+            if isinstance(t, DTensor):
+                return t.to_local()
+        except Exception:
+            pass
+        return t
+
+    @torch.no_grad()
+    def _init_shadow(self, module):
+        if self.mode == "rank0_full":
+            cpu_state = gather_state_dict_on_cpu_rank0(module, device=None)
+            if self.rank == 0:
+                self.shadow = {
+                    k: v.detach().clone().float().cpu()
+                    for k, v in cpu_state.items()
+                }
+            else:
+                self.shadow = {}
+            return
+
+        # local_shard: maintain EMA of local shards for requires_grad params
+        self.shadow = {}
+        for name, p in module.named_parameters():
+            if not p.requires_grad:
+                continue
+            local = self._to_local_tensor(p.detach())
+            self.shadow[name] = local.clone().float().cpu()
+
+    @torch.no_grad()
+    def update(self, module):
+        d = self.decay
+        if self.mode == "rank0_full":
+            if self.rank != 0:
+                return
+            cpu_state = gather_state_dict_on_cpu_rank0(module, device=None)
+            for n, v in cpu_state.items():
+                v_cpu = v.detach().float().cpu()
+                if n not in self.shadow:
+                    self.shadow[n] = v_cpu.clone()
+                else:
+                    self.shadow[n].mul_(d).add_(v_cpu, alpha=1.0 - d)
+            return
+
+        # local_shard: update local shard EMA on every rank
+        for name, p in module.named_parameters():
+            if not p.requires_grad:
+                continue
+            local = self._to_local_tensor(p.detach())
+            v_cpu = local.float().cpu()
+            if name not in self.shadow:
+                self.shadow[name] = v_cpu.clone()
+            else:
+                self.shadow[name].mul_(d).add_(v_cpu, alpha=1.0 - d)
+
+    def state_dict(self) -> dict[str, torch.Tensor]:
+        if self.mode == "rank0_full":
+            return {
+                k: v.clone()
+                for k, v in self.shadow.items()
+            } if self.rank == 0 else {}
+        return {k: v.clone() for k, v in self.shadow.items()}
+
+    def load_state_dict(self, sd: dict[str, torch.Tensor]):
+        self.shadow = {k: v.clone() for k, v in sd.items()}
+
+    @torch.no_grad()
+    def copy_to_unwrapped(self, module) -> None:
+        """
+        Copy EMA weights into a non-sharded (unwrapped) module. Intended for export/eval.
+        For mode="rank0_full", only rank 0 has the full EMA state.
+        """
+        if self.mode == "rank0_full" and self.rank != 0:
+            return
+        name_to_param = dict(module.named_parameters())
+        for n, w in self.shadow.items():
+            if n in name_to_param:
+                p = name_to_param[n]
+                p.data.copy_(w.to(dtype=p.dtype, device=p.device))
+
+    class _ApplyEMACtx:
+
+        def __init__(self, ema: "EMA_FSDP", module):
+            self.ema = ema
+            self.module = module
+            self.saved: dict[str, torch.Tensor] = {}
+
+        def __enter__(self):
+            if self.ema.mode != "local_shard":
+                raise RuntimeError(
+                    "EMA apply_to_model is only supported for mode='local_shard'"
+                )
+            with torch.no_grad():
+                for name, p in self.module.named_parameters():
+                    if not p.requires_grad:
+                        continue
+                    # Save local shard
+                    p_local = EMA_FSDP._to_local_tensor(p.detach())
+                    if p_local.numel() == 0:
+                        # Nothing to swap on this rank for this param
+                        continue
+                    self.saved[name] = p_local.clone().to(device=p_local.device,
+                                                          dtype=p_local.dtype)
+                    if name in self.ema.shadow:
+                        ema_cpu = self.ema.shadow[name]
+                        if ema_cpu.numel() != p_local.numel():
+                            # Shard shape mismatch (e.g., empty shard here), skip
+                            continue
+                        # Copy EMA shard into local param shard
+                        p_local.copy_(
+                            ema_cpu.to(dtype=p_local.dtype,
+                                       device=p_local.device))
+            return self.module
+
+        def __exit__(self, exc_type, exc, tb):
+            with torch.no_grad():
+                for name, p in self.module.named_parameters():
+                    if name in self.saved:
+                        p_local = EMA_FSDP._to_local_tensor(p.detach())
+                        if p_local.numel() == 0:
+                            continue
+                        saved_local = self.saved[name]
+                        if saved_local.numel() != p_local.numel():
+                            continue
+                        p_local.copy_(saved_local)
+            self.saved.clear()
+            return False
+
+    def apply_to_model(self, module):
+        return EMA_FSDP._ApplyEMACtx(self, module)

fastvideo/training/wan_self_forcing_distillation_pipeline.py

@@ -0,0 +1,72 @@
+# SPDX-License-Identifier: Apache-2.0
+import sys
+from copy import deepcopy
+
+from fastvideo.fastvideo_args import FastVideoArgs, TrainingArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.schedulers.scheduling_flow_match_euler_discrete import (
+    FlowMatchEulerDiscreteScheduler)
+from fastvideo.pipelines.basic.wan.wan_causal_dmd_pipeline import WanCausalDMDPipeline
+from fastvideo.training.self_forcing_distillation_pipeline import SelfForcingDistillationPipeline
+from fastvideo.utils import is_vsa_available
+
+vsa_available = is_vsa_available()
+
+logger = init_logger(__name__)
+
+
+class WanSelfForcingDistillationPipeline(SelfForcingDistillationPipeline):
+    """
+    A self-forcing distillation pipeline for Wan that uses the self-forcing methodology
+    with DMD for video generation.
+    """
+    _required_config_modules = [
+        "scheduler", "transformer", "vae", "real_score_transformer",
+        "fake_score_transformer"
+    ]
+
+    def create_training_stages(self, training_args: TrainingArgs):
+        """
+        May be used in future refactors.
+        """
+        pass
+
+    def initialize_validation_pipeline(self, training_args: TrainingArgs):
+        logger.info("Initializing validation pipeline...")
+        args_copy = deepcopy(training_args)
+
+        args_copy.inference_mode = True
+        validation_pipeline = WanCausalDMDPipeline.from_pretrained(
+            training_args.model_path,
+            args=args_copy,  # type: ignore
+            inference_mode=True,
+            loaded_modules={"transformer": self.get_module("transformer")},
+            tp_size=training_args.tp_size,
+            sp_size=training_args.sp_size,
+            num_gpus=training_args.num_gpus,
+            pin_cpu_memory=training_args.pin_cpu_memory,
+            dit_cpu_offload=True)
+
+        self.validation_pipeline = validation_pipeline
+
+
+def main(args) -> None:
+    logger.info("Starting Wan self-forcing distillation pipeline...")
+
+    pipeline = WanSelfForcingDistillationPipeline.from_pretrained(
+        args.pretrained_model_name_or_path, args=args)
+
+    args = pipeline.training_args
+    pipeline.train()
+    logger.info("Wan self-forcing distillation pipeline completed")
+
+
+if __name__ == "__main__":
+    argv = sys.argv
+    from fastvideo.fastvideo_args import TrainingArgs
+    from fastvideo.utils import FlexibleArgumentParser
+    parser = FlexibleArgumentParser()
+    parser = TrainingArgs.add_cli_args(parser)
+    parser = FastVideoArgs.add_cli_args(parser)
+    args = parser.parse_args()
+    main(args) 

fastvideo/utils.py

@@ -818,6 +818,17 @@ def is_vsa_available() -> bool:
     return importlib.util.find_spec("vsa") is not None
 
 
+@lru_cache(maxsize=1)
+def is_vmoba_available() -> bool:
+    if importlib.util.find_spec("csrc.attn.vmoba_attn.vmoba") is None:
+        return False
+    try:
+        import flash_attn
+        return flash_attn.__version__ >= "2.7.4"
+    except Exception:
+        return False
+
+
 # adapted from: https://github.com/Wan-Video/Wan2.2/blob/main/wan/utils/utils.py
 def masks_like(tensor,
                zero=False,

fastvideo/wan/README.md

@@ -0,0 +1,2 @@
+Code in this folder is modified from https://github.com/Wan-Video/Wan2.1
+Apache-2.0 License 

fastvideo/wan/__init__.py

@@ -0,0 +1,3 @@
+from . import configs, distributed, modules
+from .image2video import WanI2V
+from .text2video import WanT2V

fastvideo/wan/configs/__init__.py

@@ -0,0 +1,42 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+from .wan_t2v_14B import t2v_14B
+from .wan_t2v_1_3B import t2v_1_3B
+from .wan_i2v_14B import i2v_14B
+import copy
+import os
+
+os.environ['TOKENIZERS_PARALLELISM'] = 'false'
+
+
+# the config of t2i_14B is the same as t2v_14B
+t2i_14B = copy.deepcopy(t2v_14B)
+t2i_14B.__name__ = 'Config: Wan T2I 14B'
+
+WAN_CONFIGS = {
+    't2v-14B': t2v_14B,
+    't2v-1.3B': t2v_1_3B,
+    'i2v-14B': i2v_14B,
+    't2i-14B': t2i_14B,
+}
+
+SIZE_CONFIGS = {
+    '720*1280': (720, 1280),
+    '1280*720': (1280, 720),
+    '480*832': (480, 832),
+    '832*480': (832, 480),
+    '1024*1024': (1024, 1024),
+}
+
+MAX_AREA_CONFIGS = {
+    '720*1280': 720 * 1280,
+    '1280*720': 1280 * 720,
+    '480*832': 480 * 832,
+    '832*480': 832 * 480,
+}
+
+SUPPORTED_SIZES = {
+    't2v-14B': ('720*1280', '1280*720', '480*832', '832*480'),
+    't2v-1.3B': ('480*832', '832*480'),
+    'i2v-14B': ('720*1280', '1280*720', '480*832', '832*480'),
+    't2i-14B': tuple(SIZE_CONFIGS.keys()),
+}

fastvideo/wan/configs/shared_config.py

@@ -0,0 +1,19 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import torch
+from easydict import EasyDict
+
+# ------------------------ Wan shared config ------------------------#
+wan_shared_cfg = EasyDict()
+
+# t5
+wan_shared_cfg.t5_model = 'umt5_xxl'
+wan_shared_cfg.t5_dtype = torch.bfloat16
+wan_shared_cfg.text_len = 512
+
+# transformer
+wan_shared_cfg.param_dtype = torch.bfloat16
+
+# inference
+wan_shared_cfg.num_train_timesteps = 1000
+wan_shared_cfg.sample_fps = 16
+wan_shared_cfg.sample_neg_prompt = '色调艳丽，过曝，静态，细节模糊不清，字幕，风格，作品，画作，画面，静止，整体发灰，最差质量，低质量，JPEG压缩残留，丑陋的，残缺的，多余的手指，画得不好的手部，画得不好的脸部，畸形的，毁容的，形态畸形的肢体，手指融合，静止不动的画面，杂乱的背景，三条腿，背景人很多，倒着走'

fastvideo/wan/configs/wan_i2v_14B.py

@@ -0,0 +1,35 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import torch
+from easydict import EasyDict
+
+from .shared_config import wan_shared_cfg
+
+# ------------------------ Wan I2V 14B ------------------------#
+
+i2v_14B = EasyDict(__name__='Config: Wan I2V 14B')
+i2v_14B.update(wan_shared_cfg)
+
+i2v_14B.t5_checkpoint = 'models_t5_umt5-xxl-enc-bf16.pth'
+i2v_14B.t5_tokenizer = 'google/umt5-xxl'
+
+# clip
+i2v_14B.clip_model = 'clip_xlm_roberta_vit_h_14'
+i2v_14B.clip_dtype = torch.float16
+i2v_14B.clip_checkpoint = 'models_clip_open-clip-xlm-roberta-large-vit-huge-14.pth'
+i2v_14B.clip_tokenizer = 'xlm-roberta-large'
+
+# vae
+i2v_14B.vae_checkpoint = 'Wan2.1_VAE.pth'
+i2v_14B.vae_stride = (4, 8, 8)
+
+# transformer
+i2v_14B.patch_size = (1, 2, 2)
+i2v_14B.dim = 5120
+i2v_14B.ffn_dim = 13824
+i2v_14B.freq_dim = 256
+i2v_14B.num_heads = 40
+i2v_14B.num_layers = 40
+i2v_14B.window_size = (-1, -1)
+i2v_14B.qk_norm = True
+i2v_14B.cross_attn_norm = True
+i2v_14B.eps = 1e-6

fastvideo/wan/configs/wan_t2v_14B.py

@@ -0,0 +1,29 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+from easydict import EasyDict
+
+from .shared_config import wan_shared_cfg
+
+# ------------------------ Wan T2V 14B ------------------------#
+
+t2v_14B = EasyDict(__name__='Config: Wan T2V 14B')
+t2v_14B.update(wan_shared_cfg)
+
+# t5
+t2v_14B.t5_checkpoint = 'models_t5_umt5-xxl-enc-bf16.pth'
+t2v_14B.t5_tokenizer = 'google/umt5-xxl'
+
+# vae
+t2v_14B.vae_checkpoint = 'Wan2.1_VAE.pth'
+t2v_14B.vae_stride = (4, 8, 8)
+
+# transformer
+t2v_14B.patch_size = (1, 2, 2)
+t2v_14B.dim = 5120
+t2v_14B.ffn_dim = 13824
+t2v_14B.freq_dim = 256
+t2v_14B.num_heads = 40
+t2v_14B.num_layers = 40
+t2v_14B.window_size = (-1, -1)
+t2v_14B.qk_norm = True
+t2v_14B.cross_attn_norm = True
+t2v_14B.eps = 1e-6

fastvideo/wan/configs/wan_t2v_1_3B.py

@@ -0,0 +1,29 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+from easydict import EasyDict
+
+from .shared_config import wan_shared_cfg
+
+# ------------------------ Wan T2V 1.3B ------------------------#
+
+t2v_1_3B = EasyDict(__name__='Config: Wan T2V 1.3B')
+t2v_1_3B.update(wan_shared_cfg)
+
+# t5
+t2v_1_3B.t5_checkpoint = 'models_t5_umt5-xxl-enc-bf16.pth'
+t2v_1_3B.t5_tokenizer = 'google/umt5-xxl'
+
+# vae
+t2v_1_3B.vae_checkpoint = 'Wan2.1_VAE.pth'
+t2v_1_3B.vae_stride = (4, 8, 8)
+
+# transformer
+t2v_1_3B.patch_size = (1, 2, 2)
+t2v_1_3B.dim = 1536
+t2v_1_3B.ffn_dim = 8960
+t2v_1_3B.freq_dim = 256
+t2v_1_3B.num_heads = 12
+t2v_1_3B.num_layers = 30
+t2v_1_3B.window_size = (-1, -1)
+t2v_1_3B.qk_norm = True
+t2v_1_3B.cross_attn_norm = True
+t2v_1_3B.eps = 1e-6

fastvideo/wan/distributed/fsdp.py

@@ -0,0 +1,33 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+from functools import partial
+
+import torch
+from torch.distributed.fsdp import FullyShardedDataParallel as FSDP
+from torch.distributed.fsdp import MixedPrecision, ShardingStrategy
+from torch.distributed.fsdp.wrap import lambda_auto_wrap_policy
+
+
+def shard_model(
+    model,
+    device_id,
+    param_dtype=torch.bfloat16,
+    reduce_dtype=torch.float32,
+    buffer_dtype=torch.float32,
+    process_group=None,
+    sharding_strategy=ShardingStrategy.FULL_SHARD,
+    sync_module_states=True,
+):
+    model = FSDP(
+        module=model,
+        process_group=process_group,
+        sharding_strategy=sharding_strategy,
+        auto_wrap_policy=partial(
+            lambda_auto_wrap_policy, lambda_fn=lambda m: m in model.blocks),
+        mixed_precision=MixedPrecision(
+            param_dtype=param_dtype,
+            reduce_dtype=reduce_dtype,
+            buffer_dtype=buffer_dtype),
+        device_id=device_id,
+        use_orig_params=True,
+        sync_module_states=sync_module_states)
+    return model

fastvideo/wan/distributed/xdit_context_parallel.py

@@ -0,0 +1,192 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import torch
+import torch.cuda.amp as amp
+from xfuser.core.distributed import (get_sequence_parallel_rank,
+                                     get_sequence_parallel_world_size,
+                                     get_sp_group)
+from xfuser.core.long_ctx_attention import xFuserLongContextAttention
+
+from ..modules.model import sinusoidal_embedding_1d
+
+
+def pad_freqs(original_tensor, target_len):
+    seq_len, s1, s2 = original_tensor.shape
+    pad_size = target_len - seq_len
+    padding_tensor = torch.ones(
+        pad_size,
+        s1,
+        s2,
+        dtype=original_tensor.dtype,
+        device=original_tensor.device)
+    padded_tensor = torch.cat([original_tensor, padding_tensor], dim=0)
+    return padded_tensor
+
+
+@amp.autocast(enabled=False)
+def rope_apply(x, grid_sizes, freqs):
+    """
+    x:          [B, L, N, C].
+    grid_sizes: [B, 3].
+    freqs:      [M, C // 2].
+    """
+    s, n, c = x.size(1), x.size(2), x.size(3) // 2
+    # split freqs
+    freqs = freqs.split([c - 2 * (c // 3), c // 3, c // 3], dim=1)
+
+    # loop over samples
+    output = []
+    for i, (f, h, w) in enumerate(grid_sizes.tolist()):
+        seq_len = f * h * w
+
+        # precompute multipliers
+        x_i = torch.view_as_complex(x[i, :s].to(torch.float64).reshape(
+            s, n, -1, 2))
+        freqs_i = torch.cat([
+            freqs[0][:f].view(f, 1, 1, -1).expand(f, h, w, -1),
+            freqs[1][:h].view(1, h, 1, -1).expand(f, h, w, -1),
+            freqs[2][:w].view(1, 1, w, -1).expand(f, h, w, -1)
+        ],
+            dim=-1).reshape(seq_len, 1, -1)
+
+        # apply rotary embedding
+        sp_size = get_sequence_parallel_world_size()
+        sp_rank = get_sequence_parallel_rank()
+        freqs_i = pad_freqs(freqs_i, s * sp_size)
+        s_per_rank = s
+        freqs_i_rank = freqs_i[(sp_rank * s_per_rank):((sp_rank + 1) *
+                                                       s_per_rank), :, :]
+        x_i = torch.view_as_real(x_i * freqs_i_rank).flatten(2)
+        x_i = torch.cat([x_i, x[i, s:]])
+
+        # append to collection
+        output.append(x_i)
+    return torch.stack(output).float()
+
+
+def usp_dit_forward(
+    self,
+    x,
+    t,
+    context,
+    seq_len,
+    clip_fea=None,
+    y=None,
+):
+    """
+    x:              A list of videos each with shape [C, T, H, W].
+    t:              [B].
+    context:        A list of text embeddings each with shape [L, C].
+    """
+    if self.model_type == 'i2v':
+        assert clip_fea is not None and y is not None
+    # params
+    device = self.patch_embedding.weight.device
+    if self.freqs.device != device:
+        self.freqs = self.freqs.to(device)
+
+    if y is not None:
+        x = [torch.cat([u, v], dim=0) for u, v in zip(x, y)]
+
+    # embeddings
+    x = [self.patch_embedding(u.unsqueeze(0)) for u in x]
+    grid_sizes = torch.stack(
+        [torch.tensor(u.shape[2:], dtype=torch.long) for u in x])
+    x = [u.flatten(2).transpose(1, 2) for u in x]
+    seq_lens = torch.tensor([u.size(1) for u in x], dtype=torch.long)
+    assert seq_lens.max() <= seq_len
+    x = torch.cat([
+        torch.cat([u, u.new_zeros(1, seq_len - u.size(1), u.size(2))], dim=1)
+        for u in x
+    ])
+
+    # time embeddings
+    with amp.autocast(dtype=torch.float32):
+        e = self.time_embedding(
+            sinusoidal_embedding_1d(self.freq_dim, t).float())
+        e0 = self.time_projection(e).unflatten(1, (6, self.dim))
+        assert e.dtype == torch.float32 and e0.dtype == torch.float32
+
+    # context
+    context_lens = None
+    context = self.text_embedding(
+        torch.stack([
+            torch.cat([u, u.new_zeros(self.text_len - u.size(0), u.size(1))])
+            for u in context
+        ]))
+
+    if clip_fea is not None:
+        context_clip = self.img_emb(clip_fea)  # bs x 257 x dim
+        context = torch.concat([context_clip, context], dim=1)
+
+    # arguments
+    kwargs = dict(
+        e=e0,
+        seq_lens=seq_lens,
+        grid_sizes=grid_sizes,
+        freqs=self.freqs,
+        context=context,
+        context_lens=context_lens)
+
+    # Context Parallel
+    x = torch.chunk(
+        x, get_sequence_parallel_world_size(),
+        dim=1)[get_sequence_parallel_rank()]
+
+    for block in self.blocks:
+        x = block(x, **kwargs)
+
+    # head
+    x = self.head(x, e)
+
+    # Context Parallel
+    x = get_sp_group().all_gather(x, dim=1)
+
+    # unpatchify
+    x = self.unpatchify(x, grid_sizes)
+    return [u.float() for u in x]
+
+
+def usp_attn_forward(self,
+                     x,
+                     seq_lens,
+                     grid_sizes,
+                     freqs,
+                     dtype=torch.bfloat16):
+    b, s, n, d = *x.shape[:2], self.num_heads, self.head_dim
+    half_dtypes = (torch.float16, torch.bfloat16)
+
+    def half(x):
+        return x if x.dtype in half_dtypes else x.to(dtype)
+
+    # query, key, value function
+    def qkv_fn(x):
+        q = self.norm_q(self.q(x)).view(b, s, n, d)
+        k = self.norm_k(self.k(x)).view(b, s, n, d)
+        v = self.v(x).view(b, s, n, d)
+        return q, k, v
+
+    q, k, v = qkv_fn(x)
+    q = rope_apply(q, grid_sizes, freqs)
+    k = rope_apply(k, grid_sizes, freqs)
+
+    # TODO: We should use unpaded q,k,v for attention.
+    # k_lens = seq_lens // get_sequence_parallel_world_size()
+    # if k_lens is not None:
+    #     q = torch.cat([u[:l] for u, l in zip(q, k_lens)]).unsqueeze(0)
+    #     k = torch.cat([u[:l] for u, l in zip(k, k_lens)]).unsqueeze(0)
+    #     v = torch.cat([u[:l] for u, l in zip(v, k_lens)]).unsqueeze(0)
+
+    x = xFuserLongContextAttention()(
+        None,
+        query=half(q),
+        key=half(k),
+        value=half(v),
+        window_size=self.window_size)
+
+    # TODO: padding after attention.
+    # x = torch.cat([x, x.new_zeros(b, s - x.size(1), n, d)], dim=1)
+
+    # output
+    x = x.flatten(2)
+    x = self.o(x)
+    return x

fastvideo/wan/image2video.py

@@ -0,0 +1,347 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import gc
+import logging
+import math
+import os
+import random
+import sys
+import types
+from contextlib import contextmanager
+from functools import partial
+
+import numpy as np
+import torch
+import torch.cuda.amp as amp
+import torch.distributed as dist
+import torchvision.transforms.functional as TF
+from tqdm import tqdm
+
+from .distributed.fsdp import shard_model
+from .modules.clip import CLIPModel
+from .modules.model import WanModel
+from .modules.t5 import T5EncoderModel
+from .modules.vae import WanVAE
+from .utils.fm_solvers import (FlowDPMSolverMultistepScheduler,
+                               get_sampling_sigmas, retrieve_timesteps)
+from .utils.fm_solvers_unipc import FlowUniPCMultistepScheduler
+
+
+class WanI2V:
+
+    def __init__(
+        self,
+        config,
+        checkpoint_dir,
+        device_id=0,
+        rank=0,
+        t5_fsdp=False,
+        dit_fsdp=False,
+        use_usp=False,
+        t5_cpu=False,
+        init_on_cpu=True,
+    ):
+        r"""
+        Initializes the image-to-video generation model components.
+
+        Args:
+            config (EasyDict):
+                Object containing model parameters initialized from config.py
+            checkpoint_dir (`str`):
+                Path to directory containing model checkpoints
+            device_id (`int`,  *optional*, defaults to 0):
+                Id of target GPU device
+            rank (`int`,  *optional*, defaults to 0):
+                Process rank for distributed training
+            t5_fsdp (`bool`, *optional*, defaults to False):
+                Enable FSDP sharding for T5 model
+            dit_fsdp (`bool`, *optional*, defaults to False):
+                Enable FSDP sharding for DiT model
+            use_usp (`bool`, *optional*, defaults to False):
+                Enable distribution strategy of USP.
+            t5_cpu (`bool`, *optional*, defaults to False):
+                Whether to place T5 model on CPU. Only works without t5_fsdp.
+            init_on_cpu (`bool`, *optional*, defaults to True):
+                Enable initializing Transformer Model on CPU. Only works without FSDP or USP.
+        """
+        self.device = torch.device(f"cuda:{device_id}")
+        self.config = config
+        self.rank = rank
+        self.use_usp = use_usp
+        self.t5_cpu = t5_cpu
+
+        self.num_train_timesteps = config.num_train_timesteps
+        self.param_dtype = config.param_dtype
+
+        shard_fn = partial(shard_model, device_id=device_id)
+        self.text_encoder = T5EncoderModel(
+            text_len=config.text_len,
+            dtype=config.t5_dtype,
+            device=torch.device('cpu'),
+            checkpoint_path=os.path.join(checkpoint_dir, config.t5_checkpoint),
+            tokenizer_path=os.path.join(checkpoint_dir, config.t5_tokenizer),
+            shard_fn=shard_fn if t5_fsdp else None,
+        )
+
+        self.vae_stride = config.vae_stride
+        self.patch_size = config.patch_size
+        self.vae = WanVAE(
+            vae_pth=os.path.join(checkpoint_dir, config.vae_checkpoint),
+            device=self.device)
+
+        self.clip = CLIPModel(
+            dtype=config.clip_dtype,
+            device=self.device,
+            checkpoint_path=os.path.join(checkpoint_dir,
+                                         config.clip_checkpoint),
+            tokenizer_path=os.path.join(checkpoint_dir, config.clip_tokenizer))
+
+        logging.info(f"Creating WanModel from {checkpoint_dir}")
+        self.model = WanModel.from_pretrained(checkpoint_dir)
+        self.model.eval().requires_grad_(False)
+
+        if t5_fsdp or dit_fsdp or use_usp:
+            init_on_cpu = False
+
+        if use_usp:
+            from xfuser.core.distributed import \
+                get_sequence_parallel_world_size
+
+            from .distributed.xdit_context_parallel import (usp_attn_forward,
+                                                            usp_dit_forward)
+            for block in self.model.blocks:
+                block.self_attn.forward = types.MethodType(
+                    usp_attn_forward, block.self_attn)
+            self.model.forward = types.MethodType(usp_dit_forward, self.model)
+            self.sp_size = get_sequence_parallel_world_size()
+        else:
+            self.sp_size = 1
+
+        if dist.is_initialized():
+            dist.barrier()
+        if dit_fsdp:
+            self.model = shard_fn(self.model)
+        else:
+            if not init_on_cpu:
+                self.model.to(self.device)
+
+        self.sample_neg_prompt = config.sample_neg_prompt
+
+    def generate(self,
+                 input_prompt,
+                 img,
+                 max_area=720 * 1280,
+                 frame_num=81,
+                 shift=5.0,
+                 sample_solver='unipc',
+                 sampling_steps=40,
+                 guide_scale=5.0,
+                 n_prompt="",
+                 seed=-1,
+                 offload_model=True):
+        r"""
+        Generates video frames from input image and text prompt using diffusion process.
+
+        Args:
+            input_prompt (`str`):
+                Text prompt for content generation.
+            img (PIL.Image.Image):
+                Input image tensor. Shape: [3, H, W]
+            max_area (`int`, *optional*, defaults to 720*1280):
+                Maximum pixel area for latent space calculation. Controls video resolution scaling
+            frame_num (`int`, *optional*, defaults to 81):
+                How many frames to sample from a video. The number should be 4n+1
+            shift (`float`, *optional*, defaults to 5.0):
+                Noise schedule shift parameter. Affects temporal dynamics
+                [NOTE]: If you want to generate a 480p video, it is recommended to set the shift value to 3.0.
+            sample_solver (`str`, *optional*, defaults to 'unipc'):
+                Solver used to sample the video.
+            sampling_steps (`int`, *optional*, defaults to 40):
+                Number of diffusion sampling steps. Higher values improve quality but slow generation
+            guide_scale (`float`, *optional*, defaults 5.0):
+                Classifier-free guidance scale. Controls prompt adherence vs. creativity
+            n_prompt (`str`, *optional*, defaults to ""):
+                Negative prompt for content exclusion. If not given, use `config.sample_neg_prompt`
+            seed (`int`, *optional*, defaults to -1):
+                Random seed for noise generation. If -1, use random seed
+            offload_model (`bool`, *optional*, defaults to True):
+                If True, offloads models to CPU during generation to save VRAM
+
+        Returns:
+            torch.Tensor:
+                Generated video frames tensor. Dimensions: (C, N H, W) where:
+                - C: Color channels (3 for RGB)
+                - N: Number of frames (81)
+                - H: Frame height (from max_area)
+                - W: Frame width from max_area)
+        """
+        img = TF.to_tensor(img).sub_(0.5).div_(0.5).to(self.device)
+
+        F = frame_num
+        h, w = img.shape[1:]
+        aspect_ratio = h / w
+        lat_h = round(
+            np.sqrt(max_area * aspect_ratio) // self.vae_stride[1] //
+            self.patch_size[1] * self.patch_size[1])
+        lat_w = round(
+            np.sqrt(max_area / aspect_ratio) // self.vae_stride[2] //
+            self.patch_size[2] * self.patch_size[2])
+        h = lat_h * self.vae_stride[1]
+        w = lat_w * self.vae_stride[2]
+
+        max_seq_len = ((F - 1) // self.vae_stride[0] + 1) * lat_h * lat_w // (
+            self.patch_size[1] * self.patch_size[2])
+        max_seq_len = int(math.ceil(max_seq_len / self.sp_size)) * self.sp_size
+
+        seed = seed if seed >= 0 else random.randint(0, sys.maxsize)
+        seed_g = torch.Generator(device=self.device)
+        seed_g.manual_seed(seed)
+        noise = torch.randn(
+            16,
+            21,
+            lat_h,
+            lat_w,
+            dtype=torch.float32,
+            generator=seed_g,
+            device=self.device)
+
+        msk = torch.ones(1, 81, lat_h, lat_w, device=self.device)
+        msk[:, 1:] = 0
+        msk = torch.concat([
+            torch.repeat_interleave(msk[:, 0:1], repeats=4, dim=1), msk[:, 1:]
+        ],
+            dim=1)
+        msk = msk.view(1, msk.shape[1] // 4, 4, lat_h, lat_w)
+        msk = msk.transpose(1, 2)[0]
+
+        if n_prompt == "":
+            n_prompt = self.sample_neg_prompt
+
+        # preprocess
+        if not self.t5_cpu:
+            self.text_encoder.model.to(self.device)
+            context = self.text_encoder([input_prompt], self.device)
+            context_null = self.text_encoder([n_prompt], self.device)
+            if offload_model:
+                self.text_encoder.model.cpu()
+        else:
+            context = self.text_encoder([input_prompt], torch.device('cpu'))
+            context_null = self.text_encoder([n_prompt], torch.device('cpu'))
+            context = [t.to(self.device) for t in context]
+            context_null = [t.to(self.device) for t in context_null]
+
+        self.clip.model.to(self.device)
+        clip_context = self.clip.visual([img[:, None, :, :]])
+        if offload_model:
+            self.clip.model.cpu()
+
+        y = self.vae.encode([
+            torch.concat([
+                torch.nn.functional.interpolate(
+                    img[None].cpu(), size=(h, w), mode='bicubic').transpose(
+                        0, 1),
+                torch.zeros(3, 80, h, w)
+            ],
+                dim=1).to(self.device)
+        ])[0]
+        y = torch.concat([msk, y])
+
+        @contextmanager
+        def noop_no_sync():
+            yield
+
+        no_sync = getattr(self.model, 'no_sync', noop_no_sync)
+
+        # evaluation mode
+        with amp.autocast(dtype=self.param_dtype), torch.no_grad(), no_sync():
+
+            if sample_solver == 'unipc':
+                sample_scheduler = FlowUniPCMultistepScheduler(
+                    num_train_timesteps=self.num_train_timesteps,
+                    shift=1,
+                    use_dynamic_shifting=False)
+                sample_scheduler.set_timesteps(
+                    sampling_steps, device=self.device, shift=shift)
+                timesteps = sample_scheduler.timesteps
+            elif sample_solver == 'dpm++':
+                sample_scheduler = FlowDPMSolverMultistepScheduler(
+                    num_train_timesteps=self.num_train_timesteps,
+                    shift=1,
+                    use_dynamic_shifting=False)
+                sampling_sigmas = get_sampling_sigmas(sampling_steps, shift)
+                timesteps, _ = retrieve_timesteps(
+                    sample_scheduler,
+                    device=self.device,
+                    sigmas=sampling_sigmas)
+            else:
+                raise NotImplementedError("Unsupported solver.")
+
+            # sample videos
+            latent = noise
+
+            arg_c = {
+                'context': [context[0]],
+                'clip_fea': clip_context,
+                'seq_len': max_seq_len,
+                'y': [y],
+            }
+
+            arg_null = {
+                'context': context_null,
+                'clip_fea': clip_context,
+                'seq_len': max_seq_len,
+                'y': [y],
+            }
+
+            if offload_model:
+                torch.cuda.empty_cache()
+
+            self.model.to(self.device)
+            for _, t in enumerate(tqdm(timesteps)):
+                latent_model_input = [latent.to(self.device)]
+                timestep = [t]
+
+                timestep = torch.stack(timestep).to(self.device)
+
+                noise_pred_cond = self.model(
+                    latent_model_input, t=timestep, **arg_c)[0].to(
+                        torch.device('cpu') if offload_model else self.device)
+                if offload_model:
+                    torch.cuda.empty_cache()
+                noise_pred_uncond = self.model(
+                    latent_model_input, t=timestep, **arg_null)[0].to(
+                        torch.device('cpu') if offload_model else self.device)
+                if offload_model:
+                    torch.cuda.empty_cache()
+                noise_pred = noise_pred_uncond + guide_scale * (
+                    noise_pred_cond - noise_pred_uncond)
+
+                latent = latent.to(
+                    torch.device('cpu') if offload_model else self.device)
+
+                temp_x0 = sample_scheduler.step(
+                    noise_pred.unsqueeze(0),
+                    t,
+                    latent.unsqueeze(0),
+                    return_dict=False,
+                    generator=seed_g)[0]
+                latent = temp_x0.squeeze(0)
+
+                x0 = [latent.to(self.device)]
+                del latent_model_input, timestep
+
+            if offload_model:
+                self.model.cpu()
+                torch.cuda.empty_cache()
+
+            if self.rank == 0:
+                videos = self.vae.decode(x0)
+
+        del noise, latent
+        del sample_scheduler
+        if offload_model:
+            gc.collect()
+            torch.cuda.synchronize()
+        if dist.is_initialized():
+            dist.barrier()
+
+        return videos[0] if self.rank == 0 else None

fastvideo/wan/modules/__init__.py

@@ -0,0 +1,16 @@
+from .attention import flash_attention
+from .model import WanModel
+from .t5 import T5Decoder, T5Encoder, T5EncoderModel, T5Model
+from .tokenizers import HuggingfaceTokenizer
+from .vae import WanVAE
+
+__all__ = [
+    'WanVAE',
+    'WanModel',
+    'T5Model',
+    'T5Encoder',
+    'T5Decoder',
+    'T5EncoderModel',
+    'HuggingfaceTokenizer',
+    'flash_attention',
+]

fastvideo/wan/modules/attention.py

@@ -0,0 +1,185 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import torch
+
+try:
+    import flash_attn_interface
+
+    def is_hopper_gpu():
+        if not torch.cuda.is_available():
+            return False
+        device_name = torch.cuda.get_device_name(0).lower()
+        return "h100" in device_name or "hopper" in device_name
+    FLASH_ATTN_3_AVAILABLE = is_hopper_gpu()
+except ModuleNotFoundError:
+    FLASH_ATTN_3_AVAILABLE = False
+
+try:
+    import flash_attn
+    FLASH_ATTN_2_AVAILABLE = True
+except ModuleNotFoundError:
+    FLASH_ATTN_2_AVAILABLE = False
+
+# FLASH_ATTN_3_AVAILABLE = False
+
+import warnings
+
+__all__ = [
+    'flash_attention',
+    'attention',
+]
+
+
+def flash_attention(
+    q,
+    k,
+    v,
+    q_lens=None,
+    k_lens=None,
+    dropout_p=0.,
+    softmax_scale=None,
+    q_scale=None,
+    causal=False,
+    window_size=(-1, -1),
+    deterministic=False,
+    dtype=torch.bfloat16,
+    version=None,
+):
+    """
+    q:              [B, Lq, Nq, C1].
+    k:              [B, Lk, Nk, C1].
+    v:              [B, Lk, Nk, C2]. Nq must be divisible by Nk.
+    q_lens:         [B].
+    k_lens:         [B].
+    dropout_p:      float. Dropout probability.
+    softmax_scale:  float. The scaling of QK^T before applying softmax.
+    causal:         bool. Whether to apply causal attention mask.
+    window_size:    (left right). If not (-1, -1), apply sliding window local attention.
+    deterministic:  bool. If True, slightly slower and uses more memory.
+    dtype:          torch.dtype. Apply when dtype of q/k/v is not float16/bfloat16.
+    """
+    half_dtypes = (torch.float16, torch.bfloat16)
+    assert dtype in half_dtypes
+    assert q.device.type == 'cuda' and q.size(-1) <= 256
+
+    # params
+    b, lq, lk, out_dtype = q.size(0), q.size(1), k.size(1), q.dtype
+
+    def half(x):
+        return x if x.dtype in half_dtypes else x.to(dtype)
+
+    # preprocess query
+    if q_lens is None:
+        q = half(q.flatten(0, 1))
+        q_lens = torch.tensor(
+            [lq] * b, dtype=torch.int32).to(
+                device=q.device, non_blocking=True)
+    else:
+        q = half(torch.cat([u[:v] for u, v in zip(q, q_lens)]))
+
+    # preprocess key, value
+    if k_lens is None:
+        k = half(k.flatten(0, 1))
+        v = half(v.flatten(0, 1))
+        k_lens = torch.tensor(
+            [lk] * b, dtype=torch.int32).to(
+                device=k.device, non_blocking=True)
+    else:
+        k = half(torch.cat([u[:v] for u, v in zip(k, k_lens)]))
+        v = half(torch.cat([u[:v] for u, v in zip(v, k_lens)]))
+
+    q = q.to(v.dtype)
+    k = k.to(v.dtype)
+
+    if q_scale is not None:
+        q = q * q_scale
+
+    if version is not None and version == 3 and not FLASH_ATTN_3_AVAILABLE:
+        warnings.warn(
+            'Flash attention 3 is not available, use flash attention 2 instead.'
+        )
+
+    # apply attention
+    if (version is None or version == 3) and FLASH_ATTN_3_AVAILABLE:
+        # Note: dropout_p, window_size are not supported in FA3 now.
+        x = flash_attn_interface.flash_attn_varlen_func(
+            q=q,
+            k=k,
+            v=v,
+            cu_seqlens_q=torch.cat([q_lens.new_zeros([1]), q_lens]).cumsum(
+                0, dtype=torch.int32).to(q.device, non_blocking=True),
+            cu_seqlens_k=torch.cat([k_lens.new_zeros([1]), k_lens]).cumsum(
+                0, dtype=torch.int32).to(q.device, non_blocking=True),
+            max_seqlen_q=lq,
+            max_seqlen_k=lk,
+            softmax_scale=softmax_scale,
+            causal=causal,
+            deterministic=deterministic)[0].unflatten(0, (b, lq))
+    else:
+        assert FLASH_ATTN_2_AVAILABLE
+        x = flash_attn.flash_attn_varlen_func(
+            q=q,
+            k=k,
+            v=v,
+            cu_seqlens_q=torch.cat([q_lens.new_zeros([1]), q_lens]).cumsum(
+                0, dtype=torch.int32).to(q.device, non_blocking=True),
+            cu_seqlens_k=torch.cat([k_lens.new_zeros([1]), k_lens]).cumsum(
+                0, dtype=torch.int32).to(q.device, non_blocking=True),
+            max_seqlen_q=lq,
+            max_seqlen_k=lk,
+            dropout_p=dropout_p,
+            softmax_scale=softmax_scale,
+            causal=causal,
+            window_size=window_size,
+            deterministic=deterministic).unflatten(0, (b, lq))
+
+    # output
+    return x.type(out_dtype)
+
+
+def attention(
+    q,
+    k,
+    v,
+    q_lens=None,
+    k_lens=None,
+    dropout_p=0.,
+    softmax_scale=None,
+    q_scale=None,
+    causal=False,
+    window_size=(-1, -1),
+    deterministic=False,
+    dtype=torch.bfloat16,
+    fa_version=None,
+):
+    if FLASH_ATTN_2_AVAILABLE or FLASH_ATTN_3_AVAILABLE:
+        return flash_attention(
+            q=q,
+            k=k,
+            v=v,
+            q_lens=q_lens,
+            k_lens=k_lens,
+            dropout_p=dropout_p,
+            softmax_scale=softmax_scale,
+            q_scale=q_scale,
+            causal=causal,
+            window_size=window_size,
+            deterministic=deterministic,
+            dtype=dtype,
+            version=fa_version,
+        )
+    else:
+        if q_lens is not None or k_lens is not None:
+            warnings.warn(
+                'Padding mask is disabled when using scaled_dot_product_attention. It can have a significant impact on performance.'
+            )
+        attn_mask = None
+
+        q = q.transpose(1, 2).to(dtype)
+        k = k.transpose(1, 2).to(dtype)
+        v = v.transpose(1, 2).to(dtype)
+
+        out = torch.nn.functional.scaled_dot_product_attention(
+            q, k, v, attn_mask=attn_mask, is_causal=causal, dropout_p=dropout_p)
+
+        out = out.transpose(1, 2).contiguous()
+        return out

fastvideo/wan/modules/clip.py

@@ -0,0 +1,542 @@
+# Modified from ``https://github.com/openai/CLIP'' and ``https://github.com/mlfoundations/open_clip''
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import logging
+import math
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+import torchvision.transforms as T
+
+from .attention import flash_attention
+from .tokenizers import HuggingfaceTokenizer
+from .xlm_roberta import XLMRoberta
+
+__all__ = [
+    'XLMRobertaCLIP',
+    'clip_xlm_roberta_vit_h_14',
+    'CLIPModel',
+]
+
+
+def pos_interpolate(pos, seq_len):
+    if pos.size(1) == seq_len:
+        return pos
+    else:
+        src_grid = int(math.sqrt(pos.size(1)))
+        tar_grid = int(math.sqrt(seq_len))
+        n = pos.size(1) - src_grid * src_grid
+        return torch.cat([
+            pos[:, :n],
+            F.interpolate(
+                pos[:, n:].float().reshape(1, src_grid, src_grid, -1).permute(
+                    0, 3, 1, 2),
+                size=(tar_grid, tar_grid),
+                mode='bicubic',
+                align_corners=False).flatten(2).transpose(1, 2)
+        ],
+            dim=1)
+
+
+class QuickGELU(nn.Module):
+
+    def forward(self, x):
+        return x * torch.sigmoid(1.702 * x)
+
+
+class LayerNorm(nn.LayerNorm):
+
+    def forward(self, x):
+        return super().forward(x.float()).type_as(x)
+
+
+class SelfAttention(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 num_heads,
+                 causal=False,
+                 attn_dropout=0.0,
+                 proj_dropout=0.0):
+        assert dim % num_heads == 0
+        super().__init__()
+        self.dim = dim
+        self.num_heads = num_heads
+        self.head_dim = dim // num_heads
+        self.causal = causal
+        self.attn_dropout = attn_dropout
+        self.proj_dropout = proj_dropout
+
+        # layers
+        self.to_qkv = nn.Linear(dim, dim * 3)
+        self.proj = nn.Linear(dim, dim)
+
+    def forward(self, x):
+        """
+        x:   [B, L, C].
+        """
+        b, s, c, n, d = *x.size(), self.num_heads, self.head_dim
+
+        # compute query, key, value
+        q, k, v = self.to_qkv(x).view(b, s, 3, n, d).unbind(2)
+
+        # compute attention
+        p = self.attn_dropout if self.training else 0.0
+        x = flash_attention(q, k, v, dropout_p=p, causal=self.causal, version=2)
+        x = x.reshape(b, s, c)
+
+        # output
+        x = self.proj(x)
+        x = F.dropout(x, self.proj_dropout, self.training)
+        return x
+
+
+class SwiGLU(nn.Module):
+
+    def __init__(self, dim, mid_dim):
+        super().__init__()
+        self.dim = dim
+        self.mid_dim = mid_dim
+
+        # layers
+        self.fc1 = nn.Linear(dim, mid_dim)
+        self.fc2 = nn.Linear(dim, mid_dim)
+        self.fc3 = nn.Linear(mid_dim, dim)
+
+    def forward(self, x):
+        x = F.silu(self.fc1(x)) * self.fc2(x)
+        x = self.fc3(x)
+        return x
+
+
+class AttentionBlock(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 mlp_ratio,
+                 num_heads,
+                 post_norm=False,
+                 causal=False,
+                 activation='quick_gelu',
+                 attn_dropout=0.0,
+                 proj_dropout=0.0,
+                 norm_eps=1e-5):
+        assert activation in ['quick_gelu', 'gelu', 'swi_glu']
+        super().__init__()
+        self.dim = dim
+        self.mlp_ratio = mlp_ratio
+        self.num_heads = num_heads
+        self.post_norm = post_norm
+        self.causal = causal
+        self.norm_eps = norm_eps
+
+        # layers
+        self.norm1 = LayerNorm(dim, eps=norm_eps)
+        self.attn = SelfAttention(dim, num_heads, causal, attn_dropout,
+                                  proj_dropout)
+        self.norm2 = LayerNorm(dim, eps=norm_eps)
+        if activation == 'swi_glu':
+            self.mlp = SwiGLU(dim, int(dim * mlp_ratio))
+        else:
+            self.mlp = nn.Sequential(
+                nn.Linear(dim, int(dim * mlp_ratio)),
+                QuickGELU() if activation == 'quick_gelu' else nn.GELU(),
+                nn.Linear(int(dim * mlp_ratio), dim), nn.Dropout(proj_dropout))
+
+    def forward(self, x):
+        if self.post_norm:
+            x = x + self.norm1(self.attn(x))
+            x = x + self.norm2(self.mlp(x))
+        else:
+            x = x + self.attn(self.norm1(x))
+            x = x + self.mlp(self.norm2(x))
+        return x
+
+
+class AttentionPool(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 mlp_ratio,
+                 num_heads,
+                 activation='gelu',
+                 proj_dropout=0.0,
+                 norm_eps=1e-5):
+        assert dim % num_heads == 0
+        super().__init__()
+        self.dim = dim
+        self.mlp_ratio = mlp_ratio
+        self.num_heads = num_heads
+        self.head_dim = dim // num_heads
+        self.proj_dropout = proj_dropout
+        self.norm_eps = norm_eps
+
+        # layers
+        gain = 1.0 / math.sqrt(dim)
+        self.cls_embedding = nn.Parameter(gain * torch.randn(1, 1, dim))
+        self.to_q = nn.Linear(dim, dim)
+        self.to_kv = nn.Linear(dim, dim * 2)
+        self.proj = nn.Linear(dim, dim)
+        self.norm = LayerNorm(dim, eps=norm_eps)
+        self.mlp = nn.Sequential(
+            nn.Linear(dim, int(dim * mlp_ratio)),
+            QuickGELU() if activation == 'quick_gelu' else nn.GELU(),
+            nn.Linear(int(dim * mlp_ratio), dim), nn.Dropout(proj_dropout))
+
+    def forward(self, x):
+        """
+        x:  [B, L, C].
+        """
+        b, s, c, n, d = *x.size(), self.num_heads, self.head_dim
+
+        # compute query, key, value
+        q = self.to_q(self.cls_embedding).view(1, 1, n, d).expand(b, -1, -1, -1)
+        k, v = self.to_kv(x).view(b, s, 2, n, d).unbind(2)
+
+        # compute attention
+        x = flash_attention(q, k, v, version=2)
+        x = x.reshape(b, 1, c)
+
+        # output
+        x = self.proj(x)
+        x = F.dropout(x, self.proj_dropout, self.training)
+
+        # mlp
+        x = x + self.mlp(self.norm(x))
+        return x[:, 0]
+
+
+class VisionTransformer(nn.Module):
+
+    def __init__(self,
+                 image_size=224,
+                 patch_size=16,
+                 dim=768,
+                 mlp_ratio=4,
+                 out_dim=512,
+                 num_heads=12,
+                 num_layers=12,
+                 pool_type='token',
+                 pre_norm=True,
+                 post_norm=False,
+                 activation='quick_gelu',
+                 attn_dropout=0.0,
+                 proj_dropout=0.0,
+                 embedding_dropout=0.0,
+                 norm_eps=1e-5):
+        if image_size % patch_size != 0:
+            print(
+                '[WARNING] image_size is not divisible by patch_size',
+                flush=True)
+        assert pool_type in ('token', 'token_fc', 'attn_pool')
+        out_dim = out_dim or dim
+        super().__init__()
+        self.image_size = image_size
+        self.patch_size = patch_size
+        self.num_patches = (image_size // patch_size)**2
+        self.dim = dim
+        self.mlp_ratio = mlp_ratio
+        self.out_dim = out_dim
+        self.num_heads = num_heads
+        self.num_layers = num_layers
+        self.pool_type = pool_type
+        self.post_norm = post_norm
+        self.norm_eps = norm_eps
+
+        # embeddings
+        gain = 1.0 / math.sqrt(dim)
+        self.patch_embedding = nn.Conv2d(
+            3,
+            dim,
+            kernel_size=patch_size,
+            stride=patch_size,
+            bias=not pre_norm)
+        if pool_type in ('token', 'token_fc'):
+            self.cls_embedding = nn.Parameter(gain * torch.randn(1, 1, dim))
+        self.pos_embedding = nn.Parameter(gain * torch.randn(
+            1, self.num_patches +
+            (1 if pool_type in ('token', 'token_fc') else 0), dim))
+        self.dropout = nn.Dropout(embedding_dropout)
+
+        # transformer
+        self.pre_norm = LayerNorm(dim, eps=norm_eps) if pre_norm else None
+        self.transformer = nn.Sequential(*[
+            AttentionBlock(dim, mlp_ratio, num_heads, post_norm, False,
+                           activation, attn_dropout, proj_dropout, norm_eps)
+            for _ in range(num_layers)
+        ])
+        self.post_norm = LayerNorm(dim, eps=norm_eps)
+
+        # head
+        if pool_type == 'token':
+            self.head = nn.Parameter(gain * torch.randn(dim, out_dim))
+        elif pool_type == 'token_fc':
+            self.head = nn.Linear(dim, out_dim)
+        elif pool_type == 'attn_pool':
+            self.head = AttentionPool(dim, mlp_ratio, num_heads, activation,
+                                      proj_dropout, norm_eps)
+
+    def forward(self, x, interpolation=False, use_31_block=False):
+        b = x.size(0)
+
+        # embeddings
+        x = self.patch_embedding(x).flatten(2).permute(0, 2, 1)
+        if self.pool_type in ('token', 'token_fc'):
+            x = torch.cat([self.cls_embedding.expand(b, -1, -1), x], dim=1)
+        if interpolation:
+            e = pos_interpolate(self.pos_embedding, x.size(1))
+        else:
+            e = self.pos_embedding
+        x = self.dropout(x + e)
+        if self.pre_norm is not None:
+            x = self.pre_norm(x)
+
+        # transformer
+        if use_31_block:
+            x = self.transformer[:-1](x)
+            return x
+        else:
+            x = self.transformer(x)
+            return x
+
+
+class XLMRobertaWithHead(XLMRoberta):
+
+    def __init__(self, **kwargs):
+        self.out_dim = kwargs.pop('out_dim')
+        super().__init__(**kwargs)
+
+        # head
+        mid_dim = (self.dim + self.out_dim) // 2
+        self.head = nn.Sequential(
+            nn.Linear(self.dim, mid_dim, bias=False), nn.GELU(),
+            nn.Linear(mid_dim, self.out_dim, bias=False))
+
+    def forward(self, ids):
+        # xlm-roberta
+        x = super().forward(ids)
+
+        # average pooling
+        mask = ids.ne(self.pad_id).unsqueeze(-1).to(x)
+        x = (x * mask).sum(dim=1) / mask.sum(dim=1)
+
+        # head
+        x = self.head(x)
+        return x
+
+
+class XLMRobertaCLIP(nn.Module):
+
+    def __init__(self,
+                 embed_dim=1024,
+                 image_size=224,
+                 patch_size=14,
+                 vision_dim=1280,
+                 vision_mlp_ratio=4,
+                 vision_heads=16,
+                 vision_layers=32,
+                 vision_pool='token',
+                 vision_pre_norm=True,
+                 vision_post_norm=False,
+                 activation='gelu',
+                 vocab_size=250002,
+                 max_text_len=514,
+                 type_size=1,
+                 pad_id=1,
+                 text_dim=1024,
+                 text_heads=16,
+                 text_layers=24,
+                 text_post_norm=True,
+                 text_dropout=0.1,
+                 attn_dropout=0.0,
+                 proj_dropout=0.0,
+                 embedding_dropout=0.0,
+                 norm_eps=1e-5):
+        super().__init__()
+        self.embed_dim = embed_dim
+        self.image_size = image_size
+        self.patch_size = patch_size
+        self.vision_dim = vision_dim
+        self.vision_mlp_ratio = vision_mlp_ratio
+        self.vision_heads = vision_heads
+        self.vision_layers = vision_layers
+        self.vision_pre_norm = vision_pre_norm
+        self.vision_post_norm = vision_post_norm
+        self.activation = activation
+        self.vocab_size = vocab_size
+        self.max_text_len = max_text_len
+        self.type_size = type_size
+        self.pad_id = pad_id
+        self.text_dim = text_dim
+        self.text_heads = text_heads
+        self.text_layers = text_layers
+        self.text_post_norm = text_post_norm
+        self.norm_eps = norm_eps
+
+        # models
+        self.visual = VisionTransformer(
+            image_size=image_size,
+            patch_size=patch_size,
+            dim=vision_dim,
+            mlp_ratio=vision_mlp_ratio,
+            out_dim=embed_dim,
+            num_heads=vision_heads,
+            num_layers=vision_layers,
+            pool_type=vision_pool,
+            pre_norm=vision_pre_norm,
+            post_norm=vision_post_norm,
+            activation=activation,
+            attn_dropout=attn_dropout,
+            proj_dropout=proj_dropout,
+            embedding_dropout=embedding_dropout,
+            norm_eps=norm_eps)
+        self.textual = XLMRobertaWithHead(
+            vocab_size=vocab_size,
+            max_seq_len=max_text_len,
+            type_size=type_size,
+            pad_id=pad_id,
+            dim=text_dim,
+            out_dim=embed_dim,
+            num_heads=text_heads,
+            num_layers=text_layers,
+            post_norm=text_post_norm,
+            dropout=text_dropout)
+        self.log_scale = nn.Parameter(math.log(1 / 0.07) * torch.ones([]))
+
+    def forward(self, imgs, txt_ids):
+        """
+        imgs:       [B, 3, H, W] of torch.float32.
+        - mean:     [0.48145466, 0.4578275, 0.40821073]
+        - std:      [0.26862954, 0.26130258, 0.27577711]
+        txt_ids:    [B, L] of torch.long.
+                    Encoded by data.CLIPTokenizer.
+        """
+        xi = self.visual(imgs)
+        xt = self.textual(txt_ids)
+        return xi, xt
+
+    def param_groups(self):
+        groups = [{
+            'params': [
+                p for n, p in self.named_parameters()
+                if 'norm' in n or n.endswith('bias')
+            ],
+            'weight_decay': 0.0
+        }, {
+            'params': [
+                p for n, p in self.named_parameters()
+                if not ('norm' in n or n.endswith('bias'))
+            ]
+        }]
+        return groups
+
+
+def _clip(pretrained=False,
+          pretrained_name=None,
+          model_cls=XLMRobertaCLIP,
+          return_transforms=False,
+          return_tokenizer=False,
+          tokenizer_padding='eos',
+          dtype=torch.float32,
+          device='cpu',
+          **kwargs):
+    # init a model on device
+    with torch.device(device):
+        model = model_cls(**kwargs)
+
+    # set device
+    model = model.to(dtype=dtype, device=device)
+    output = (model,)
+
+    # init transforms
+    if return_transforms:
+        # mean and std
+        if 'siglip' in pretrained_name.lower():
+            mean, std = [0.5, 0.5, 0.5], [0.5, 0.5, 0.5]
+        else:
+            mean = [0.48145466, 0.4578275, 0.40821073]
+            std = [0.26862954, 0.26130258, 0.27577711]
+
+        # transforms
+        transforms = T.Compose([
+            T.Resize((model.image_size, model.image_size),
+                     interpolation=T.InterpolationMode.BICUBIC),
+            T.ToTensor(),
+            T.Normalize(mean=mean, std=std)
+        ])
+        output += (transforms,)
+    return output[0] if len(output) == 1 else output
+
+
+def clip_xlm_roberta_vit_h_14(
+        pretrained=False,
+        pretrained_name='open-clip-xlm-roberta-large-vit-huge-14',
+        **kwargs):
+    cfg = dict(
+        embed_dim=1024,
+        image_size=224,
+        patch_size=14,
+        vision_dim=1280,
+        vision_mlp_ratio=4,
+        vision_heads=16,
+        vision_layers=32,
+        vision_pool='token',
+        activation='gelu',
+        vocab_size=250002,
+        max_text_len=514,
+        type_size=1,
+        pad_id=1,
+        text_dim=1024,
+        text_heads=16,
+        text_layers=24,
+        text_post_norm=True,
+        text_dropout=0.1,
+        attn_dropout=0.0,
+        proj_dropout=0.0,
+        embedding_dropout=0.0)
+    cfg.update(**kwargs)
+    return _clip(pretrained, pretrained_name, XLMRobertaCLIP, **cfg)
+
+
+class CLIPModel:
+
+    def __init__(self, dtype, device, checkpoint_path, tokenizer_path):
+        self.dtype = dtype
+        self.device = device
+        self.checkpoint_path = checkpoint_path
+        self.tokenizer_path = tokenizer_path
+
+        # init model
+        self.model, self.transforms = clip_xlm_roberta_vit_h_14(
+            pretrained=False,
+            return_transforms=True,
+            return_tokenizer=False,
+            dtype=dtype,
+            device=device)
+        self.model = self.model.eval().requires_grad_(False)
+        logging.info(f'loading {checkpoint_path}')
+        self.model.load_state_dict(
+            torch.load(checkpoint_path, map_location='cpu'))
+
+        # init tokenizer
+        self.tokenizer = HuggingfaceTokenizer(
+            name=tokenizer_path,
+            seq_len=self.model.max_text_len - 2,
+            clean='whitespace')
+
+    def visual(self, videos):
+        # preprocess
+        size = (self.model.image_size,) * 2
+        videos = torch.cat([
+            F.interpolate(
+                u.transpose(0, 1),
+                size=size,
+                mode='bicubic',
+                align_corners=False) for u in videos
+        ])
+        videos = self.transforms.transforms[-1](videos.mul_(0.5).add_(0.5))
+
+        # forward
+        with torch.cuda.amp.autocast(dtype=self.dtype):
+            out = self.model.visual(videos, use_31_block=True)
+            return out

fastvideo/wan/modules/model.py

@@ -0,0 +1,934 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import math
+
+import torch
+import torch.nn as nn
+from diffusers.configuration_utils import ConfigMixin, register_to_config
+from diffusers.models.modeling_utils import ModelMixin
+from einops import repeat
+
+from .attention import flash_attention
+
+__all__ = ['WanModel']
+
+
+def sinusoidal_embedding_1d(dim, position):
+    # preprocess
+    assert dim % 2 == 0
+    half = dim // 2
+    position = position.type(torch.float64)
+
+    # calculation
+    sinusoid = torch.outer(
+        position, torch.pow(10000, -torch.arange(half).to(position).div(half)))
+    x = torch.cat([torch.cos(sinusoid), torch.sin(sinusoid)], dim=1)
+    return x
+
+
+# @amp.autocast(enabled=False)
+def rope_params(max_seq_len, dim, theta=10000):
+    assert dim % 2 == 0
+    freqs = torch.outer(
+        torch.arange(max_seq_len),
+        1.0 / torch.pow(theta,
+                        torch.arange(0, dim, 2).to(torch.float64).div(dim)))
+    freqs = torch.polar(torch.ones_like(freqs), freqs)
+    return freqs
+
+
+# @amp.autocast(enabled=False)
+def rope_apply(x, grid_sizes, freqs):
+    n, c = x.size(2), x.size(3) // 2
+
+    # split freqs
+    freqs = freqs.split([c - 2 * (c // 3), c // 3, c // 3], dim=1)
+
+    # loop over samples
+    output = []
+    for i, (f, h, w) in enumerate(grid_sizes.tolist()):
+        seq_len = f * h * w
+
+        # precompute multipliers
+        x_i = torch.view_as_complex(x[i, :seq_len].to(torch.float64).reshape(
+            seq_len, n, -1, 2))
+        freqs_i = torch.cat([
+            freqs[0][:f].view(f, 1, 1, -1).expand(f, h, w, -1),
+            freqs[1][:h].view(1, h, 1, -1).expand(f, h, w, -1),
+            freqs[2][:w].view(1, 1, w, -1).expand(f, h, w, -1)
+        ],
+            dim=-1).reshape(seq_len, 1, -1)
+
+        # apply rotary embedding
+        x_i = torch.view_as_real(x_i * freqs_i).flatten(2)
+        x_i = torch.cat([x_i, x[i, seq_len:]])
+
+        # append to collection
+        output.append(x_i)
+    return torch.stack(output).type_as(x)
+
+
+class WanRMSNorm(nn.Module):
+
+    def __init__(self, dim, eps=1e-5):
+        super().__init__()
+        self.dim = dim
+        self.eps = eps
+        self.weight = nn.Parameter(torch.ones(dim))
+
+    def forward(self, x):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L, C]
+        """
+        return self._norm(x.float()).type_as(x) * self.weight
+
+    def _norm(self, x):
+        return x * torch.rsqrt(x.pow(2).mean(dim=-1, keepdim=True) + self.eps)
+
+
+class WanLayerNorm(nn.LayerNorm):
+
+    def __init__(self, dim, eps=1e-6, elementwise_affine=False):
+        super().__init__(dim, elementwise_affine=elementwise_affine, eps=eps)
+
+    def forward(self, x):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L, C]
+        """
+        return super().forward(x).type_as(x)
+
+
+class WanSelfAttention(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 num_heads,
+                 window_size=(-1, -1),
+                 qk_norm=True,
+                 eps=1e-6):
+        assert dim % num_heads == 0
+        super().__init__()
+        self.dim = dim
+        self.num_heads = num_heads
+        self.head_dim = dim // num_heads
+        self.window_size = window_size
+        self.qk_norm = qk_norm
+        self.eps = eps
+
+        # layers
+        self.q = nn.Linear(dim, dim)
+        self.k = nn.Linear(dim, dim)
+        self.v = nn.Linear(dim, dim)
+        self.o = nn.Linear(dim, dim)
+        self.norm_q = WanRMSNorm(dim, eps=eps) if qk_norm else nn.Identity()
+        self.norm_k = WanRMSNorm(dim, eps=eps) if qk_norm else nn.Identity()
+
+    def forward(self, x, seq_lens, grid_sizes, freqs):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L, num_heads, C / num_heads]
+            seq_lens(Tensor): Shape [B]
+            grid_sizes(Tensor): Shape [B, 3], the second dimension contains (F, H, W)
+            freqs(Tensor): Rope freqs, shape [1024, C / num_heads / 2]
+        """
+        b, s, n, d = *x.shape[:2], self.num_heads, self.head_dim
+
+        # query, key, value function
+        def qkv_fn(x):
+            q = self.norm_q(self.q(x)).view(b, s, n, d)
+            k = self.norm_k(self.k(x)).view(b, s, n, d)
+            v = self.v(x).view(b, s, n, d)
+            return q, k, v
+
+        q, k, v = qkv_fn(x)
+
+        print(f"query sum: {torch.sum(q.float()).item()}")
+
+        q = rope_apply(q, grid_sizes, freqs)
+
+        print(f"query after rotary embeddings sum: {torch.sum(q.float()).item()}")
+
+        x = flash_attention(
+            q=q,
+            k=rope_apply(k, grid_sizes, freqs),
+            v=v,
+            k_lens=seq_lens,
+            window_size=self.window_size)
+
+        # output
+        x = x.flatten(2)
+        x = self.o(x)
+
+        print(f"attn_output sum: {torch.sum(x.float()).item()}")
+        return x
+
+
+class WanT2VCrossAttention(WanSelfAttention):
+
+    def forward(self, x, context, context_lens, crossattn_cache=None):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L1, C]
+            context(Tensor): Shape [B, L2, C]
+            context_lens(Tensor): Shape [B]
+            crossattn_cache (List[dict], *optional*): Contains the cached key and value tensors for context embedding.
+        """
+        b, n, d = x.size(0), self.num_heads, self.head_dim
+
+        # compute query, key, value
+        q = self.norm_q(self.q(x)).view(b, -1, n, d)
+
+        if crossattn_cache is not None:
+            if not crossattn_cache["is_init"]:
+                crossattn_cache["is_init"] = True
+                k = self.norm_k(self.k(context)).view(b, -1, n, d)
+                v = self.v(context).view(b, -1, n, d)
+                crossattn_cache["k"] = k
+                crossattn_cache["v"] = v
+            else:
+                k = crossattn_cache["k"]
+                v = crossattn_cache["v"]
+        else:
+            k = self.norm_k(self.k(context)).view(b, -1, n, d)
+            v = self.v(context).view(b, -1, n, d)
+
+        # compute attention
+        x = flash_attention(q, k, v, k_lens=context_lens)
+
+        # output
+        x = x.flatten(2)
+        x = self.o(x)
+        return x
+
+
+class WanGanCrossAttention(WanSelfAttention):
+
+    def forward(self, x, context, crossattn_cache=None):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L1, C]
+            context(Tensor): Shape [B, L2, C]
+            context_lens(Tensor): Shape [B]
+            crossattn_cache (List[dict], *optional*): Contains the cached key and value tensors for context embedding.
+        """
+        b, n, d = x.size(0), self.num_heads, self.head_dim
+
+        # compute query, key, value
+        qq = self.norm_q(self.q(context)).view(b, 1, -1, d)
+
+        kk = self.norm_k(self.k(x)).view(b, -1, n, d)
+        vv = self.v(x).view(b, -1, n, d)
+
+        # compute attention
+        x = flash_attention(qq, kk, vv)
+
+        # output
+        x = x.flatten(2)
+        x = self.o(x)
+        return x
+
+
+class WanI2VCrossAttention(WanSelfAttention):
+
+    def __init__(self,
+                 dim,
+                 num_heads,
+                 window_size=(-1, -1),
+                 qk_norm=True,
+                 eps=1e-6):
+        super().__init__(dim, num_heads, window_size, qk_norm, eps)
+
+        self.k_img = nn.Linear(dim, dim)
+        self.v_img = nn.Linear(dim, dim)
+        # self.alpha = nn.Parameter(torch.zeros((1, )))
+        self.norm_k_img = WanRMSNorm(
+            dim, eps=eps) if qk_norm else nn.Identity()
+
+    def forward(self, x, context, context_lens):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L1, C]
+            context(Tensor): Shape [B, L2, C]
+            context_lens(Tensor): Shape [B]
+        """
+        context_img = context[:, :257]
+        context = context[:, 257:]
+        b, n, d = x.size(0), self.num_heads, self.head_dim
+
+        # compute query, key, value
+        q = self.norm_q(self.q(x)).view(b, -1, n, d)
+        k = self.norm_k(self.k(context)).view(b, -1, n, d)
+        v = self.v(context).view(b, -1, n, d)
+        k_img = self.norm_k_img(self.k_img(context_img)).view(b, -1, n, d)
+        v_img = self.v_img(context_img).view(b, -1, n, d)
+        img_x = flash_attention(q, k_img, v_img, k_lens=None)
+        # compute attention
+        x = flash_attention(q, k, v, k_lens=context_lens)
+
+        # output
+        x = x.flatten(2)
+        img_x = img_x.flatten(2)
+        x = x + img_x
+        x = self.o(x)
+        return x
+
+
+WAN_CROSSATTENTION_CLASSES = {
+    't2v_cross_attn': WanT2VCrossAttention,
+    'i2v_cross_attn': WanI2VCrossAttention,
+}
+
+
+class WanAttentionBlock(nn.Module):
+
+    def __init__(self,
+                 cross_attn_type,
+                 dim,
+                 ffn_dim,
+                 num_heads,
+                 window_size=(-1, -1),
+                 qk_norm=True,
+                 cross_attn_norm=False,
+                 eps=1e-6):
+        super().__init__()
+        self.dim = dim
+        self.ffn_dim = ffn_dim
+        self.num_heads = num_heads
+        self.window_size = window_size
+        self.qk_norm = qk_norm
+        self.cross_attn_norm = cross_attn_norm
+        self.eps = eps
+
+        # layers
+        self.norm1 = WanLayerNorm(dim, eps)
+        self.self_attn = WanSelfAttention(dim, num_heads, window_size, qk_norm,
+                                          eps)
+        self.norm3 = WanLayerNorm(
+            dim, eps,
+            elementwise_affine=True) if cross_attn_norm else nn.Identity()
+        self.cross_attn = WAN_CROSSATTENTION_CLASSES[cross_attn_type](dim,
+                                                                      num_heads,
+                                                                      (-1, -1),
+                                                                      qk_norm,
+                                                                      eps)
+        self.norm2 = WanLayerNorm(dim, eps)
+        self.ffn = nn.Sequential(
+            nn.Linear(dim, ffn_dim), nn.GELU(approximate='tanh'),
+            nn.Linear(ffn_dim, dim))
+
+        # modulation
+        self.modulation = nn.Parameter(torch.randn(1, 6, dim) / dim**0.5)
+
+    def forward(
+        self,
+        x,
+        e,
+        seq_lens,
+        grid_sizes,
+        freqs,
+        context,
+        context_lens,
+    ):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L, C]
+            e(Tensor): Shape [B, 6, C]
+            seq_lens(Tensor): Shape [B], length of each sequence in batch
+            grid_sizes(Tensor): Shape [B, 3], the second dimension contains (F, H, W)
+            freqs(Tensor): Rope freqs, shape [1024, C / num_heads / 2]
+        """
+        # assert e.dtype == torch.float32
+        # with amp.autocast(dtype=torch.float32):
+        e = (self.modulation + e).chunk(6, dim=1)
+        # assert e[0].dtype == torch.float32
+
+
+        norm_x = self.norm1(x) * (1 + e[1]) + e[0]
+        print(f"norm_hidden_states sum: {torch.sum(norm_x.float()).item()}")
+        # self-attention
+        y = self.self_attn(
+            norm_x, seq_lens, grid_sizes,
+            freqs)
+        # with amp.autocast(dtype=torch.float32):
+        x = x + y * e[2]
+
+        # cross-attention & ffn function
+        def cross_attn_ffn(x, context, context_lens, e):
+            x = x + self.cross_attn(self.norm3(x), context, context_lens)
+            y = self.ffn(self.norm2(x) * (1 + e[4]) + e[3])
+            # with amp.autocast(dtype=torch.float32):
+            x = x + y * e[5]
+            return x
+
+        x = cross_attn_ffn(x, context, context_lens, e)
+        return x
+
+
+class GanAttentionBlock(nn.Module):
+
+    def __init__(self,
+                 dim=1536,
+                 ffn_dim=8192,
+                 num_heads=12,
+                 window_size=(-1, -1),
+                 qk_norm=True,
+                 cross_attn_norm=True,
+                 eps=1e-6):
+        super().__init__()
+        self.dim = dim
+        self.ffn_dim = ffn_dim
+        self.num_heads = num_heads
+        self.window_size = window_size
+        self.qk_norm = qk_norm
+        self.cross_attn_norm = cross_attn_norm
+        self.eps = eps
+
+        # layers
+        # self.norm1 = WanLayerNorm(dim, eps)
+        # self.self_attn = WanSelfAttention(dim, num_heads, window_size, qk_norm,
+        #   eps)
+        self.norm3 = WanLayerNorm(
+            dim, eps,
+            elementwise_affine=True) if cross_attn_norm else nn.Identity()
+
+        self.norm2 = WanLayerNorm(dim, eps)
+        self.ffn = nn.Sequential(
+            nn.Linear(dim, ffn_dim), nn.GELU(approximate='tanh'),
+            nn.Linear(ffn_dim, dim))
+
+        self.cross_attn = WanGanCrossAttention(dim, num_heads,
+                                               (-1, -1),
+                                               qk_norm,
+                                               eps)
+
+        # modulation
+        # self.modulation = nn.Parameter(torch.randn(1, 6, dim) / dim**0.5)
+
+    def forward(
+        self,
+        x,
+        context,
+        # seq_lens,
+        # grid_sizes,
+        # freqs,
+        # context,
+        # context_lens,
+    ):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L, C]
+            e(Tensor): Shape [B, 6, C]
+            seq_lens(Tensor): Shape [B], length of each sequence in batch
+            grid_sizes(Tensor): Shape [B, 3], the second dimension contains (F, H, W)
+            freqs(Tensor): Rope freqs, shape [1024, C / num_heads / 2]
+        """
+        # assert e.dtype == torch.float32
+        # with amp.autocast(dtype=torch.float32):
+        # e = (self.modulation + e).chunk(6, dim=1)
+        # assert e[0].dtype == torch.float32
+
+        # # self-attention
+        # y = self.self_attn(
+        #     self.norm1(x) * (1 + e[1]) + e[0], seq_lens, grid_sizes,
+        #     freqs)
+        # # with amp.autocast(dtype=torch.float32):
+        # x = x + y * e[2]
+
+        # cross-attention & ffn function
+        def cross_attn_ffn(x, context):
+            token = context + self.cross_attn(self.norm3(x), context)
+            y = self.ffn(self.norm2(token)) + token  # * (1 + e[4]) + e[3])
+            # with amp.autocast(dtype=torch.float32):
+            # x = x + y * e[5]
+            return y
+
+        x = cross_attn_ffn(x, context)
+        return x
+
+
+class Head(nn.Module):
+
+    def __init__(self, dim, out_dim, patch_size, eps=1e-6):
+        super().__init__()
+        self.dim = dim
+        self.out_dim = out_dim
+        self.patch_size = patch_size
+        self.eps = eps
+
+        # layers
+        out_dim = math.prod(patch_size) * out_dim
+        self.norm = WanLayerNorm(dim, eps)
+        self.head = nn.Linear(dim, out_dim)
+
+        # modulation
+        self.modulation = nn.Parameter(torch.randn(1, 2, dim) / dim**0.5)
+
+    def forward(self, x, e):
+        r"""
+        Args:
+            x(Tensor): Shape [B, L1, C]
+            e(Tensor): Shape [B, C]
+        """
+        # assert e.dtype == torch.float32
+        # with amp.autocast(dtype=torch.float32):
+        e = (self.modulation + e.unsqueeze(1)).chunk(2, dim=1)
+        x = (self.head(self.norm(x) * (1 + e[1]) + e[0]))
+        return x
+
+
+class MLPProj(torch.nn.Module):
+
+    def __init__(self, in_dim, out_dim):
+        super().__init__()
+
+        self.proj = torch.nn.Sequential(
+            torch.nn.LayerNorm(in_dim), torch.nn.Linear(in_dim, in_dim),
+            torch.nn.GELU(), torch.nn.Linear(in_dim, out_dim),
+            torch.nn.LayerNorm(out_dim))
+
+    def forward(self, image_embeds):
+        clip_extra_context_tokens = self.proj(image_embeds)
+        return clip_extra_context_tokens
+
+
+class RegisterTokens(nn.Module):
+    def __init__(self, num_registers: int, dim: int):
+        super().__init__()
+        self.register_tokens = nn.Parameter(torch.randn(num_registers, dim) * 0.02)
+        self.rms_norm = WanRMSNorm(dim, eps=1e-6)
+
+    def forward(self):
+        return self.rms_norm(self.register_tokens)
+
+    def reset_parameters(self):
+        nn.init.normal_(self.register_tokens, std=0.02)
+
+
+class WanModel(ModelMixin, ConfigMixin):
+    r"""
+    Wan diffusion backbone supporting both text-to-video and image-to-video.
+    """
+
+    ignore_for_config = [
+        'patch_size', 'cross_attn_norm', 'qk_norm', 'text_dim', 'window_size'
+    ]
+    _no_split_modules = ['WanAttentionBlock']
+    _supports_gradient_checkpointing = True
+
+    @register_to_config
+    def __init__(self,
+                 model_type='t2v',
+                 patch_size=(1, 2, 2),
+                 text_len=512,
+                 in_dim=16,
+                 dim=2048,
+                 ffn_dim=8192,
+                 freq_dim=256,
+                 text_dim=4096,
+                 out_dim=16,
+                 num_heads=16,
+                 num_layers=32,
+                 window_size=(-1, -1),
+                 qk_norm=True,
+                 cross_attn_norm=True,
+                 eps=1e-6):
+        r"""
+        Initialize the diffusion model backbone.
+
+        Args:
+            model_type (`str`, *optional*, defaults to 't2v'):
+                Model variant - 't2v' (text-to-video) or 'i2v' (image-to-video)
+            patch_size (`tuple`, *optional*, defaults to (1, 2, 2)):
+                3D patch dimensions for video embedding (t_patch, h_patch, w_patch)
+            text_len (`int`, *optional*, defaults to 512):
+                Fixed length for text embeddings
+            in_dim (`int`, *optional*, defaults to 16):
+                Input video channels (C_in)
+            dim (`int`, *optional*, defaults to 2048):
+                Hidden dimension of the transformer
+            ffn_dim (`int`, *optional*, defaults to 8192):
+                Intermediate dimension in feed-forward network
+            freq_dim (`int`, *optional*, defaults to 256):
+                Dimension for sinusoidal time embeddings
+            text_dim (`int`, *optional*, defaults to 4096):
+                Input dimension for text embeddings
+            out_dim (`int`, *optional*, defaults to 16):
+                Output video channels (C_out)
+            num_heads (`int`, *optional*, defaults to 16):
+                Number of attention heads
+            num_layers (`int`, *optional*, defaults to 32):
+                Number of transformer blocks
+            window_size (`tuple`, *optional*, defaults to (-1, -1)):
+                Window size for local attention (-1 indicates global attention)
+            qk_norm (`bool`, *optional*, defaults to True):
+                Enable query/key normalization
+            cross_attn_norm (`bool`, *optional*, defaults to False):
+                Enable cross-attention normalization
+            eps (`float`, *optional*, defaults to 1e-6):
+                Epsilon value for normalization layers
+        """
+
+        super().__init__()
+
+        assert model_type in ['t2v', 'i2v']
+        self.model_type = model_type
+
+        self.patch_size = patch_size
+        self.text_len = text_len
+        self.in_dim = in_dim
+        self.dim = dim
+        self.ffn_dim = ffn_dim
+        self.freq_dim = freq_dim
+        self.text_dim = text_dim
+        self.out_dim = out_dim
+        self.num_heads = num_heads
+        self.num_layers = num_layers
+        self.window_size = window_size
+        self.qk_norm = qk_norm
+        self.cross_attn_norm = cross_attn_norm
+        self.eps = eps
+        self.local_attn_size = 21
+
+        # embeddings
+        self.patch_embedding = nn.Conv3d(
+            in_dim, dim, kernel_size=patch_size, stride=patch_size)
+        self.text_embedding = nn.Sequential(
+            nn.Linear(text_dim, dim), nn.GELU(approximate='tanh'),
+            nn.Linear(dim, dim))
+
+        self.time_embedding = nn.Sequential(
+            nn.Linear(freq_dim, dim), nn.SiLU(), nn.Linear(dim, dim))
+        self.time_projection = nn.Sequential(
+            nn.SiLU(), nn.Linear(dim, dim * 6))
+
+        # blocks
+        cross_attn_type = 't2v_cross_attn' if model_type == 't2v' else 'i2v_cross_attn'
+        self.blocks = nn.ModuleList([
+            WanAttentionBlock(cross_attn_type, dim, ffn_dim, num_heads,
+                              window_size, qk_norm, cross_attn_norm, eps)
+            for _ in range(num_layers)
+        ])
+
+        # head
+        self.head = Head(dim, out_dim, patch_size, eps)
+
+        # buffers (don't use register_buffer otherwise dtype will be changed in to())
+        assert (dim % num_heads) == 0 and (dim // num_heads) % 2 == 0
+        d = dim // num_heads
+        self.freqs = torch.cat([
+            rope_params(1024, d - 4 * (d // 6)),
+            rope_params(1024, 2 * (d // 6)),
+            rope_params(1024, 2 * (d // 6))
+        ],
+            dim=1)
+
+        if model_type == 'i2v':
+            self.img_emb = MLPProj(1280, dim)
+
+        # initialize weights
+        self.init_weights()
+
+        self.gradient_checkpointing = False
+
+    def _set_gradient_checkpointing(self, module, value=False):
+        self.gradient_checkpointing = value
+
+    def forward(
+        self,
+        *args,
+        **kwargs
+    ):
+        # if kwargs.get('classify_mode', False) is True:
+        # kwargs.pop('classify_mode')
+        # return self._forward_classify(*args, **kwargs)
+        # else:
+        return self._forward(*args, **kwargs)
+
+    def _forward(
+        self,
+        x,
+        t,
+        context,
+        seq_len,
+        classify_mode=False,
+        concat_time_embeddings=False,
+        register_tokens=None,
+        cls_pred_branch=None,
+        gan_ca_blocks=None,
+        clip_fea=None,
+        y=None,
+    ):
+        r"""
+        Forward pass through the diffusion model
+
+        Args:
+            x (List[Tensor]):
+                List of input video tensors, each with shape [C_in, F, H, W]
+            t (Tensor):
+                Diffusion timesteps tensor of shape [B]
+            context (List[Tensor]):
+                List of text embeddings each with shape [L, C]
+            seq_len (`int`):
+                Maximum sequence length for positional encoding
+            clip_fea (Tensor, *optional*):
+                CLIP image features for image-to-video mode
+            y (List[Tensor], *optional*):
+                Conditional video inputs for image-to-video mode, same shape as x
+
+        Returns:
+            List[Tensor]:
+                List of denoised video tensors with original input shapes [C_out, F, H / 8, W / 8]
+        """
+        if self.model_type == 'i2v':
+            assert clip_fea is not None and y is not None
+        # params
+        device = self.patch_embedding.weight.device
+        if self.freqs.device != device:
+            self.freqs = self.freqs.to(device)
+
+        if y is not None:
+            x = [torch.cat([u, v], dim=0) for u, v in zip(x, y)]
+
+        # embeddings
+        x = [self.patch_embedding(u.unsqueeze(0)) for u in x]
+        grid_sizes = torch.stack(
+            [torch.tensor(u.shape[2:], dtype=torch.long) for u in x])
+        x = [u.flatten(2).transpose(1, 2) for u in x]
+        seq_lens = torch.tensor([u.size(1) for u in x], dtype=torch.long)
+        assert seq_lens.max() <= seq_len
+        x = torch.cat([
+            torch.cat([u, u.new_zeros(1, seq_len - u.size(1), u.size(2))],
+                      dim=1) for u in x
+        ])
+
+        # time embeddings
+        # with amp.autocast(dtype=torch.float32):
+        e = self.time_embedding(
+            sinusoidal_embedding_1d(self.freq_dim, t).type_as(x))
+        e0 = self.time_projection(e).unflatten(1, (6, self.dim))
+        # assert e.dtype == torch.float32 and e0.dtype == torch.float32
+
+        # context
+        context_lens = None
+        context = self.text_embedding(
+            torch.stack([
+                torch.cat(
+                    [u, u.new_zeros(self.text_len - u.size(0), u.size(1))])
+                for u in context
+            ]))
+
+        if clip_fea is not None:
+            context_clip = self.img_emb(clip_fea)  # bs x 257 x dim
+            context = torch.concat([context_clip, context], dim=1)
+
+        # arguments
+        kwargs = dict(
+            e=e0,
+            seq_lens=seq_lens,
+            grid_sizes=grid_sizes,
+            freqs=self.freqs,
+            context=context,
+            context_lens=context_lens)
+
+        def create_custom_forward(module):
+            def custom_forward(*inputs, **kwargs):
+                return module(*inputs, **kwargs)
+            return custom_forward
+
+        # TODO: Tune the number of blocks for feature extraction
+        final_x = None
+        if classify_mode:
+            assert register_tokens is not None
+            assert gan_ca_blocks is not None
+            assert cls_pred_branch is not None
+
+            final_x = []
+            registers = repeat(register_tokens(), "n d -> b n d", b=x.shape[0])
+            # x = torch.cat([registers, x], dim=1)
+
+        gan_idx = 0
+        for ii, block in enumerate(self.blocks):
+            if torch.is_grad_enabled() and self.gradient_checkpointing:
+                x = torch.utils.checkpoint.checkpoint(
+                    create_custom_forward(block),
+                    x, **kwargs,
+                    use_reentrant=False,
+                )
+            else:
+                x = block(x, **kwargs)
+
+            if classify_mode and ii in [13, 21, 29]:
+                gan_token = registers[:, gan_idx: gan_idx + 1]
+                final_x.append(gan_ca_blocks[gan_idx](x, gan_token))
+                gan_idx += 1
+
+        if classify_mode:
+            final_x = torch.cat(final_x, dim=1)
+            if concat_time_embeddings:
+                final_x = cls_pred_branch(torch.cat([final_x, 10 * e[:, None, :]], dim=1).view(final_x.shape[0], -1))
+            else:
+                final_x = cls_pred_branch(final_x.view(final_x.shape[0], -1))
+
+        # head
+        x = self.head(x, e)
+
+        # unpatchify
+        x = self.unpatchify(x, grid_sizes)
+
+        if classify_mode:
+            return torch.stack(x), final_x
+
+        return torch.stack(x)
+
+    def _forward_classify(
+        self,
+        x,
+        t,
+        context,
+        seq_len,
+        register_tokens,
+        cls_pred_branch,
+        clip_fea=None,
+        y=None,
+    ):
+        r"""
+        Feature extraction through the diffusion model
+
+        Args:
+            x (List[Tensor]):
+                List of input video tensors, each with shape [C_in, F, H, W]
+            t (Tensor):
+                Diffusion timesteps tensor of shape [B]
+            context (List[Tensor]):
+                List of text embeddings each with shape [L, C]
+            seq_len (`int`):
+                Maximum sequence length for positional encoding
+            clip_fea (Tensor, *optional*):
+                CLIP image features for image-to-video mode
+            y (List[Tensor], *optional*):
+                Conditional video inputs for image-to-video mode, same shape as x
+
+        Returns:
+            List[Tensor]:
+                List of video features with original input shapes [C_block, F, H / 8, W / 8]
+        """
+        if self.model_type == 'i2v':
+            assert clip_fea is not None and y is not None
+        # params
+        device = self.patch_embedding.weight.device
+        if self.freqs.device != device:
+            self.freqs = self.freqs.to(device)
+
+        if y is not None:
+            x = [torch.cat([u, v], dim=0) for u, v in zip(x, y)]
+
+        # embeddings
+        x = [self.patch_embedding(u.unsqueeze(0)) for u in x]
+        grid_sizes = torch.stack(
+            [torch.tensor(u.shape[2:], dtype=torch.long) for u in x])
+        x = [u.flatten(2).transpose(1, 2) for u in x]
+        seq_lens = torch.tensor([u.size(1) for u in x], dtype=torch.long)
+        assert seq_lens.max() <= seq_len
+        x = torch.cat([
+            torch.cat([u, u.new_zeros(1, seq_len - u.size(1), u.size(2))],
+                      dim=1) for u in x
+        ])
+
+        # time embeddings
+        # with amp.autocast(dtype=torch.float32):
+        e = self.time_embedding(
+            sinusoidal_embedding_1d(self.freq_dim, t).type_as(x))
+        e0 = self.time_projection(e).unflatten(1, (6, self.dim))
+        # assert e.dtype == torch.float32 and e0.dtype == torch.float32
+
+        # context
+        context_lens = None
+        context = self.text_embedding(
+            torch.stack([
+                torch.cat(
+                    [u, u.new_zeros(self.text_len - u.size(0), u.size(1))])
+                for u in context
+            ]))
+
+        if clip_fea is not None:
+            context_clip = self.img_emb(clip_fea)  # bs x 257 x dim
+            context = torch.concat([context_clip, context], dim=1)
+
+        # arguments
+        kwargs = dict(
+            e=e0,
+            seq_lens=seq_lens,
+            grid_sizes=grid_sizes,
+            freqs=self.freqs,
+            context=context,
+            context_lens=context_lens)
+
+        def create_custom_forward(module):
+            def custom_forward(*inputs, **kwargs):
+                return module(*inputs, **kwargs)
+            return custom_forward
+
+        # TODO: Tune the number of blocks for feature extraction
+        for block in self.blocks[:16]:
+            if torch.is_grad_enabled() and self.gradient_checkpointing:
+                x = torch.utils.checkpoint.checkpoint(
+                    create_custom_forward(block),
+                    x, **kwargs,
+                    use_reentrant=False,
+                )
+            else:
+                x = block(x, **kwargs)
+
+        # unpatchify
+        x = self.unpatchify(x, grid_sizes, c=self.dim // 4)
+        return torch.stack(x)
+
+    def unpatchify(self, x, grid_sizes, c=None):
+        r"""
+        Reconstruct video tensors from patch embeddings.
+
+        Args:
+            x (List[Tensor]):
+                List of patchified features, each with shape [L, C_out * prod(patch_size)]
+            grid_sizes (Tensor):
+                Original spatial-temporal grid dimensions before patching,
+                    shape [B, 3] (3 dimensions correspond to F_patches, H_patches, W_patches)
+
+        Returns:
+            List[Tensor]:
+                Reconstructed video tensors with shape [C_out, F, H / 8, W / 8]
+        """
+
+        c = self.out_dim if c is None else c
+        out = []
+        for u, v in zip(x, grid_sizes.tolist()):
+            u = u[:math.prod(v)].view(*v, *self.patch_size, c)
+            u = torch.einsum('fhwpqrc->cfphqwr', u)
+            u = u.reshape(c, *[i * j for i, j in zip(v, self.patch_size)])
+            out.append(u)
+        return out
+
+    def init_weights(self):
+        r"""
+        Initialize model parameters using Xavier initialization.
+        """
+
+        # basic init
+        for m in self.modules():
+            if isinstance(m, nn.Linear):
+                nn.init.xavier_uniform_(m.weight)
+                if m.bias is not None:
+                    nn.init.zeros_(m.bias)
+
+        # init embeddings
+        nn.init.xavier_uniform_(self.patch_embedding.weight.flatten(1))
+        for m in self.text_embedding.modules():
+            if isinstance(m, nn.Linear):
+                nn.init.normal_(m.weight, std=.02)
+        for m in self.time_embedding.modules():
+            if isinstance(m, nn.Linear):
+                nn.init.normal_(m.weight, std=.02)
+
+        # init output layer
+        nn.init.zeros_(self.head.head.weight)

fastvideo/wan/modules/t5.py

@@ -0,0 +1,513 @@
+# Modified from transformers.models.t5.modeling_t5
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import logging
+import math
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+from .tokenizers import HuggingfaceTokenizer
+
+__all__ = [
+    'T5Model',
+    'T5Encoder',
+    'T5Decoder',
+    'T5EncoderModel',
+]
+
+
+def fp16_clamp(x):
+    if x.dtype == torch.float16 and torch.isinf(x).any():
+        clamp = torch.finfo(x.dtype).max - 1000
+        x = torch.clamp(x, min=-clamp, max=clamp)
+    return x
+
+
+def init_weights(m):
+    if isinstance(m, T5LayerNorm):
+        nn.init.ones_(m.weight)
+    elif isinstance(m, T5Model):
+        nn.init.normal_(m.token_embedding.weight, std=1.0)
+    elif isinstance(m, T5FeedForward):
+        nn.init.normal_(m.gate[0].weight, std=m.dim**-0.5)
+        nn.init.normal_(m.fc1.weight, std=m.dim**-0.5)
+        nn.init.normal_(m.fc2.weight, std=m.dim_ffn**-0.5)
+    elif isinstance(m, T5Attention):
+        nn.init.normal_(m.q.weight, std=(m.dim * m.dim_attn)**-0.5)
+        nn.init.normal_(m.k.weight, std=m.dim**-0.5)
+        nn.init.normal_(m.v.weight, std=m.dim**-0.5)
+        nn.init.normal_(m.o.weight, std=(m.num_heads * m.dim_attn)**-0.5)
+    elif isinstance(m, T5RelativeEmbedding):
+        nn.init.normal_(
+            m.embedding.weight, std=(2 * m.num_buckets * m.num_heads)**-0.5)
+
+
+class GELU(nn.Module):
+
+    def forward(self, x):
+        return 0.5 * x * (1.0 + torch.tanh(
+            math.sqrt(2.0 / math.pi) * (x + 0.044715 * torch.pow(x, 3.0))))
+
+
+class T5LayerNorm(nn.Module):
+
+    def __init__(self, dim, eps=1e-6):
+        super(T5LayerNorm, self).__init__()
+        self.dim = dim
+        self.eps = eps
+        self.weight = nn.Parameter(torch.ones(dim))
+
+    def forward(self, x):
+        x = x * torch.rsqrt(x.float().pow(2).mean(dim=-1, keepdim=True) +
+                            self.eps)
+        if self.weight.dtype in [torch.float16, torch.bfloat16]:
+            x = x.type_as(self.weight)
+        return self.weight * x
+
+
+class T5Attention(nn.Module):
+
+    def __init__(self, dim, dim_attn, num_heads, dropout=0.1):
+        assert dim_attn % num_heads == 0
+        super(T5Attention, self).__init__()
+        self.dim = dim
+        self.dim_attn = dim_attn
+        self.num_heads = num_heads
+        self.head_dim = dim_attn // num_heads
+
+        # layers
+        self.q = nn.Linear(dim, dim_attn, bias=False)
+        self.k = nn.Linear(dim, dim_attn, bias=False)
+        self.v = nn.Linear(dim, dim_attn, bias=False)
+        self.o = nn.Linear(dim_attn, dim, bias=False)
+        self.dropout = nn.Dropout(dropout)
+
+    def forward(self, x, context=None, mask=None, pos_bias=None):
+        """
+        x:          [B, L1, C].
+        context:    [B, L2, C] or None.
+        mask:       [B, L2] or [B, L1, L2] or None.
+        """
+        # check inputs
+        context = x if context is None else context
+        b, n, c = x.size(0), self.num_heads, self.head_dim
+
+        # compute query, key, value
+        q = self.q(x).view(b, -1, n, c)
+        k = self.k(context).view(b, -1, n, c)
+        v = self.v(context).view(b, -1, n, c)
+
+        # attention bias
+        attn_bias = x.new_zeros(b, n, q.size(1), k.size(1))
+        if pos_bias is not None:
+            attn_bias += pos_bias
+        if mask is not None:
+            assert mask.ndim in [2, 3]
+            mask = mask.view(b, 1, 1,
+                             -1) if mask.ndim == 2 else mask.unsqueeze(1)
+            attn_bias.masked_fill_(mask == 0, torch.finfo(x.dtype).min)
+
+        # compute attention (T5 does not use scaling)
+        attn = torch.einsum('binc,bjnc->bnij', q, k) + attn_bias
+        attn = F.softmax(attn.float(), dim=-1).type_as(attn)
+        x = torch.einsum('bnij,bjnc->binc', attn, v)
+
+        # output
+        x = x.reshape(b, -1, n * c)
+        x = self.o(x)
+        x = self.dropout(x)
+        return x
+
+
+class T5FeedForward(nn.Module):
+
+    def __init__(self, dim, dim_ffn, dropout=0.1):
+        super(T5FeedForward, self).__init__()
+        self.dim = dim
+        self.dim_ffn = dim_ffn
+
+        # layers
+        self.gate = nn.Sequential(nn.Linear(dim, dim_ffn, bias=False), GELU())
+        self.fc1 = nn.Linear(dim, dim_ffn, bias=False)
+        self.fc2 = nn.Linear(dim_ffn, dim, bias=False)
+        self.dropout = nn.Dropout(dropout)
+
+    def forward(self, x):
+        x = self.fc1(x) * self.gate(x)
+        x = self.dropout(x)
+        x = self.fc2(x)
+        x = self.dropout(x)
+        return x
+
+
+class T5SelfAttention(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 dim_attn,
+                 dim_ffn,
+                 num_heads,
+                 num_buckets,
+                 shared_pos=True,
+                 dropout=0.1):
+        super(T5SelfAttention, self).__init__()
+        self.dim = dim
+        self.dim_attn = dim_attn
+        self.dim_ffn = dim_ffn
+        self.num_heads = num_heads
+        self.num_buckets = num_buckets
+        self.shared_pos = shared_pos
+
+        # layers
+        self.norm1 = T5LayerNorm(dim)
+        self.attn = T5Attention(dim, dim_attn, num_heads, dropout)
+        self.norm2 = T5LayerNorm(dim)
+        self.ffn = T5FeedForward(dim, dim_ffn, dropout)
+        self.pos_embedding = None if shared_pos else T5RelativeEmbedding(
+            num_buckets, num_heads, bidirectional=True)
+
+    def forward(self, x, mask=None, pos_bias=None):
+        e = pos_bias if self.shared_pos else self.pos_embedding(
+            x.size(1), x.size(1))
+        x = fp16_clamp(x + self.attn(self.norm1(x), mask=mask, pos_bias=e))
+        x = fp16_clamp(x + self.ffn(self.norm2(x)))
+        return x
+
+
+class T5CrossAttention(nn.Module):
+
+    def __init__(self,
+                 dim,
+                 dim_attn,
+                 dim_ffn,
+                 num_heads,
+                 num_buckets,
+                 shared_pos=True,
+                 dropout=0.1):
+        super(T5CrossAttention, self).__init__()
+        self.dim = dim
+        self.dim_attn = dim_attn
+        self.dim_ffn = dim_ffn
+        self.num_heads = num_heads
+        self.num_buckets = num_buckets
+        self.shared_pos = shared_pos
+
+        # layers
+        self.norm1 = T5LayerNorm(dim)
+        self.self_attn = T5Attention(dim, dim_attn, num_heads, dropout)
+        self.norm2 = T5LayerNorm(dim)
+        self.cross_attn = T5Attention(dim, dim_attn, num_heads, dropout)
+        self.norm3 = T5LayerNorm(dim)
+        self.ffn = T5FeedForward(dim, dim_ffn, dropout)
+        self.pos_embedding = None if shared_pos else T5RelativeEmbedding(
+            num_buckets, num_heads, bidirectional=False)
+
+    def forward(self,
+                x,
+                mask=None,
+                encoder_states=None,
+                encoder_mask=None,
+                pos_bias=None):
+        e = pos_bias if self.shared_pos else self.pos_embedding(
+            x.size(1), x.size(1))
+        x = fp16_clamp(x + self.self_attn(self.norm1(x), mask=mask, pos_bias=e))
+        x = fp16_clamp(x + self.cross_attn(
+            self.norm2(x), context=encoder_states, mask=encoder_mask))
+        x = fp16_clamp(x + self.ffn(self.norm3(x)))
+        return x
+
+
+class T5RelativeEmbedding(nn.Module):
+
+    def __init__(self, num_buckets, num_heads, bidirectional, max_dist=128):
+        super(T5RelativeEmbedding, self).__init__()
+        self.num_buckets = num_buckets
+        self.num_heads = num_heads
+        self.bidirectional = bidirectional
+        self.max_dist = max_dist
+
+        # layers
+        self.embedding = nn.Embedding(num_buckets, num_heads)
+
+    def forward(self, lq, lk):
+        device = self.embedding.weight.device
+        # rel_pos = torch.arange(lk).unsqueeze(0).to(device) - \
+        #     torch.arange(lq).unsqueeze(1).to(device)
+        rel_pos = torch.arange(lk, device=device).unsqueeze(0) - \
+            torch.arange(lq, device=device).unsqueeze(1)
+        rel_pos = self._relative_position_bucket(rel_pos)
+        rel_pos_embeds = self.embedding(rel_pos)
+        rel_pos_embeds = rel_pos_embeds.permute(2, 0, 1).unsqueeze(
+            0)  # [1, N, Lq, Lk]
+        return rel_pos_embeds.contiguous()
+
+    def _relative_position_bucket(self, rel_pos):
+        # preprocess
+        if self.bidirectional:
+            num_buckets = self.num_buckets // 2
+            rel_buckets = (rel_pos > 0).long() * num_buckets
+            rel_pos = torch.abs(rel_pos)
+        else:
+            num_buckets = self.num_buckets
+            rel_buckets = 0
+            rel_pos = -torch.min(rel_pos, torch.zeros_like(rel_pos))
+
+        # embeddings for small and large positions
+        max_exact = num_buckets // 2
+        rel_pos_large = max_exact + (torch.log(rel_pos.float() / max_exact) /
+                                     math.log(self.max_dist / max_exact) *
+                                     (num_buckets - max_exact)).long()
+        rel_pos_large = torch.min(
+            rel_pos_large, torch.full_like(rel_pos_large, num_buckets - 1))
+        rel_buckets += torch.where(rel_pos < max_exact, rel_pos, rel_pos_large)
+        return rel_buckets
+
+
+class T5Encoder(nn.Module):
+
+    def __init__(self,
+                 vocab,
+                 dim,
+                 dim_attn,
+                 dim_ffn,
+                 num_heads,
+                 num_layers,
+                 num_buckets,
+                 shared_pos=True,
+                 dropout=0.1):
+        super(T5Encoder, self).__init__()
+        self.dim = dim
+        self.dim_attn = dim_attn
+        self.dim_ffn = dim_ffn
+        self.num_heads = num_heads
+        self.num_layers = num_layers
+        self.num_buckets = num_buckets
+        self.shared_pos = shared_pos
+
+        # layers
+        self.token_embedding = vocab if isinstance(vocab, nn.Embedding) \
+            else nn.Embedding(vocab, dim)
+        self.pos_embedding = T5RelativeEmbedding(
+            num_buckets, num_heads, bidirectional=True) if shared_pos else None
+        self.dropout = nn.Dropout(dropout)
+        self.blocks = nn.ModuleList([
+            T5SelfAttention(dim, dim_attn, dim_ffn, num_heads, num_buckets,
+                            shared_pos, dropout) for _ in range(num_layers)
+        ])
+        self.norm = T5LayerNorm(dim)
+
+        # initialize weights
+        self.apply(init_weights)
+
+    def forward(self, ids, mask=None):
+        x = self.token_embedding(ids)
+        x = self.dropout(x)
+        e = self.pos_embedding(x.size(1),
+                               x.size(1)) if self.shared_pos else None
+        for block in self.blocks:
+            x = block(x, mask, pos_bias=e)
+        x = self.norm(x)
+        x = self.dropout(x)
+        return x
+
+
+class T5Decoder(nn.Module):
+
+    def __init__(self,
+                 vocab,
+                 dim,
+                 dim_attn,
+                 dim_ffn,
+                 num_heads,
+                 num_layers,
+                 num_buckets,
+                 shared_pos=True,
+                 dropout=0.1):
+        super(T5Decoder, self).__init__()
+        self.dim = dim
+        self.dim_attn = dim_attn
+        self.dim_ffn = dim_ffn
+        self.num_heads = num_heads
+        self.num_layers = num_layers
+        self.num_buckets = num_buckets
+        self.shared_pos = shared_pos
+
+        # layers
+        self.token_embedding = vocab if isinstance(vocab, nn.Embedding) \
+            else nn.Embedding(vocab, dim)
+        self.pos_embedding = T5RelativeEmbedding(
+            num_buckets, num_heads, bidirectional=False) if shared_pos else None
+        self.dropout = nn.Dropout(dropout)
+        self.blocks = nn.ModuleList([
+            T5CrossAttention(dim, dim_attn, dim_ffn, num_heads, num_buckets,
+                             shared_pos, dropout) for _ in range(num_layers)
+        ])
+        self.norm = T5LayerNorm(dim)
+
+        # initialize weights
+        self.apply(init_weights)
+
+    def forward(self, ids, mask=None, encoder_states=None, encoder_mask=None):
+        b, s = ids.size()
+
+        # causal mask
+        if mask is None:
+            mask = torch.tril(torch.ones(1, s, s).to(ids.device))
+        elif mask.ndim == 2:
+            mask = torch.tril(mask.unsqueeze(1).expand(-1, s, -1))
+
+        # layers
+        x = self.token_embedding(ids)
+        x = self.dropout(x)
+        e = self.pos_embedding(x.size(1),
+                               x.size(1)) if self.shared_pos else None
+        for block in self.blocks:
+            x = block(x, mask, encoder_states, encoder_mask, pos_bias=e)
+        x = self.norm(x)
+        x = self.dropout(x)
+        return x
+
+
+class T5Model(nn.Module):
+
+    def __init__(self,
+                 vocab_size,
+                 dim,
+                 dim_attn,
+                 dim_ffn,
+                 num_heads,
+                 encoder_layers,
+                 decoder_layers,
+                 num_buckets,
+                 shared_pos=True,
+                 dropout=0.1):
+        super(T5Model, self).__init__()
+        self.vocab_size = vocab_size
+        self.dim = dim
+        self.dim_attn = dim_attn
+        self.dim_ffn = dim_ffn
+        self.num_heads = num_heads
+        self.encoder_layers = encoder_layers
+        self.decoder_layers = decoder_layers
+        self.num_buckets = num_buckets
+
+        # layers
+        self.token_embedding = nn.Embedding(vocab_size, dim)
+        self.encoder = T5Encoder(self.token_embedding, dim, dim_attn, dim_ffn,
+                                 num_heads, encoder_layers, num_buckets,
+                                 shared_pos, dropout)
+        self.decoder = T5Decoder(self.token_embedding, dim, dim_attn, dim_ffn,
+                                 num_heads, decoder_layers, num_buckets,
+                                 shared_pos, dropout)
+        self.head = nn.Linear(dim, vocab_size, bias=False)
+
+        # initialize weights
+        self.apply(init_weights)
+
+    def forward(self, encoder_ids, encoder_mask, decoder_ids, decoder_mask):
+        x = self.encoder(encoder_ids, encoder_mask)
+        x = self.decoder(decoder_ids, decoder_mask, x, encoder_mask)
+        x = self.head(x)
+        return x
+
+
+def _t5(name,
+        encoder_only=False,
+        decoder_only=False,
+        return_tokenizer=False,
+        tokenizer_kwargs={},
+        dtype=torch.float32,
+        device='cpu',
+        **kwargs):
+    # sanity check
+    assert not (encoder_only and decoder_only)
+
+    # params
+    if encoder_only:
+        model_cls = T5Encoder
+        kwargs['vocab'] = kwargs.pop('vocab_size')
+        kwargs['num_layers'] = kwargs.pop('encoder_layers')
+        _ = kwargs.pop('decoder_layers')
+    elif decoder_only:
+        model_cls = T5Decoder
+        kwargs['vocab'] = kwargs.pop('vocab_size')
+        kwargs['num_layers'] = kwargs.pop('decoder_layers')
+        _ = kwargs.pop('encoder_layers')
+    else:
+        model_cls = T5Model
+
+    # init model
+    with torch.device(device):
+        model = model_cls(**kwargs)
+
+    # set device
+    model = model.to(dtype=dtype, device=device)
+
+    # init tokenizer
+    if return_tokenizer:
+        from .tokenizers import HuggingfaceTokenizer
+        tokenizer = HuggingfaceTokenizer(f'google/{name}', **tokenizer_kwargs)
+        return model, tokenizer
+    else:
+        return model
+
+
+def umt5_xxl(**kwargs):
+    cfg = dict(
+        vocab_size=256384,
+        dim=4096,
+        dim_attn=4096,
+        dim_ffn=10240,
+        num_heads=64,
+        encoder_layers=24,
+        decoder_layers=24,
+        num_buckets=32,
+        shared_pos=False,
+        dropout=0.1)
+    cfg.update(**kwargs)
+    return _t5('umt5-xxl', **cfg)
+
+
+class T5EncoderModel:
+
+    def __init__(
+        self,
+        text_len,
+        dtype=torch.bfloat16,
+        device=torch.cuda.current_device(),
+        checkpoint_path=None,
+        tokenizer_path=None,
+        shard_fn=None,
+    ):
+        self.text_len = text_len
+        self.dtype = dtype
+        self.device = device
+        self.checkpoint_path = checkpoint_path
+        self.tokenizer_path = tokenizer_path
+
+        # init model
+        model = umt5_xxl(
+            encoder_only=True,
+            return_tokenizer=False,
+            dtype=dtype,
+            device=device).eval().requires_grad_(False)
+        logging.info(f'loading {checkpoint_path}')
+        model.load_state_dict(torch.load(checkpoint_path, map_location='cpu'))
+        self.model = model
+        if shard_fn is not None:
+            self.model = shard_fn(self.model, sync_module_states=False)
+        else:
+            self.model.to(self.device)
+        # init tokenizer
+        self.tokenizer = HuggingfaceTokenizer(
+            name=tokenizer_path, seq_len=text_len, clean='whitespace')
+
+    def __call__(self, texts, device):
+        ids, mask = self.tokenizer(
+            texts, return_mask=True, add_special_tokens=True)
+        ids = ids.to(device)
+        mask = mask.to(device)
+        seq_lens = mask.gt(0).sum(dim=1).long()
+        context = self.model(ids, mask)
+        return [u[:v] for u, v in zip(context, seq_lens)]

fastvideo/wan/modules/tokenizers.py

@@ -0,0 +1,82 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import html
+import string
+
+import ftfy
+import regex as re
+from transformers import AutoTokenizer
+
+__all__ = ['HuggingfaceTokenizer']
+
+
+def basic_clean(text):
+    text = ftfy.fix_text(text)
+    text = html.unescape(html.unescape(text))
+    return text.strip()
+
+
+def whitespace_clean(text):
+    text = re.sub(r'\s+', ' ', text)
+    text = text.strip()
+    return text
+
+
+def canonicalize(text, keep_punctuation_exact_string=None):
+    text = text.replace('_', ' ')
+    if keep_punctuation_exact_string:
+        text = keep_punctuation_exact_string.join(
+            part.translate(str.maketrans('', '', string.punctuation))
+            for part in text.split(keep_punctuation_exact_string))
+    else:
+        text = text.translate(str.maketrans('', '', string.punctuation))
+    text = text.lower()
+    text = re.sub(r'\s+', ' ', text)
+    return text.strip()
+
+
+class HuggingfaceTokenizer:
+
+    def __init__(self, name, seq_len=None, clean=None, **kwargs):
+        assert clean in (None, 'whitespace', 'lower', 'canonicalize')
+        self.name = name
+        self.seq_len = seq_len
+        self.clean = clean
+
+        # init tokenizer
+        self.tokenizer = AutoTokenizer.from_pretrained(name, **kwargs)
+        self.vocab_size = self.tokenizer.vocab_size
+
+    def __call__(self, sequence, **kwargs):
+        return_mask = kwargs.pop('return_mask', False)
+
+        # arguments
+        _kwargs = {'return_tensors': 'pt'}
+        if self.seq_len is not None:
+            _kwargs.update({
+                'padding': 'max_length',
+                'truncation': True,
+                'max_length': self.seq_len
+            })
+        _kwargs.update(**kwargs)
+
+        # tokenization
+        if isinstance(sequence, str):
+            sequence = [sequence]
+        if self.clean:
+            sequence = [self._clean(u) for u in sequence]
+        ids = self.tokenizer(sequence, **_kwargs)
+
+        # output
+        if return_mask:
+            return ids.input_ids, ids.attention_mask
+        else:
+            return ids.input_ids
+
+    def _clean(self, text):
+        if self.clean == 'whitespace':
+            text = whitespace_clean(basic_clean(text))
+        elif self.clean == 'lower':
+            text = whitespace_clean(basic_clean(text)).lower()
+        elif self.clean == 'canonicalize':
+            text = canonicalize(basic_clean(text))
+        return text

fastvideo/wan/modules/vae.py

@@ -0,0 +1,683 @@
+# Copyright 2024-2025 The Alibaba Wan Team Authors. All rights reserved.
+import logging
+
+import torch
+import torch.cuda.amp as amp
+import torch.nn as nn
+import torch.nn.functional as F
+from einops import rearrange
+
+__all__ = [
+    'WanVAE',
+]
+
+CACHE_T = 2
+
+
+class CausalConv3d(nn.Conv3d):
+    """
+    Causal 3d convolusion.
+    """
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+        self._padding = (self.padding[2], self.padding[2], self.padding[1],
+                         self.padding[1], 2 * self.padding[0], 0)
+        self.padding = (0, 0, 0)
+
+    def forward(self, x, cache_x=None):
+        padding = list(self._padding)
+        if cache_x is not None and self._padding[4] > 0:
+            cache_x = cache_x.to(x.device)
+            x = torch.cat([cache_x, x], dim=2)
+            padding[4] -= cache_x.shape[2]
+        x = F.pad(x, padding)
+
+        return super().forward(x)
+
+
+class RMS_norm(nn.Module):
+
+    def __init__(self, dim, channel_first=True, images=True, bias=False):
+        super().__init__()
+        broadcastable_dims = (1, 1, 1) if not images else (1, 1)
+        shape = (dim, *broadcastable_dims) if channel_first else (dim,)
+
+        self.channel_first = channel_first
+        self.scale = dim**0.5
+        self.gamma = nn.Parameter(torch.ones(shape))
+        self.bias = nn.Parameter(torch.zeros(shape)) if bias else 0.
+
+    def forward(self, x):
+        return F.normalize(
+            x, dim=(1 if self.channel_first else
+                    -1)) * self.scale * self.gamma + self.bias
+
+
+class Upsample(nn.Upsample):
+
+    def forward(self, x):
+        """
+        Fix bfloat16 support for nearest neighbor interpolation.
+        """
+        return super().forward(x.float()).type_as(x)
+
+
+class Resample(nn.Module):
+
+    def __init__(self, dim, mode):
+        assert mode in ('none', 'upsample2d', 'upsample3d', 'downsample2d',
+                        'downsample3d')
+        super().__init__()
+        self.dim = dim
+        self.mode = mode
+
+        # layers
+        if mode == 'upsample2d':
+            self.resample = nn.Sequential(
+                Upsample(scale_factor=(2., 2.), mode='nearest'),
+                nn.Conv2d(dim, dim // 2, 3, padding=1))
+        elif mode == 'upsample3d':
+            self.resample = nn.Sequential(
+                Upsample(scale_factor=(2., 2.), mode='nearest'),
+                nn.Conv2d(dim, dim // 2, 3, padding=1))
+            self.time_conv = CausalConv3d(
+                dim, dim * 2, (3, 1, 1), padding=(1, 0, 0))
+
+        elif mode == 'downsample2d':
+            self.resample = nn.Sequential(
+                nn.ZeroPad2d((0, 1, 0, 1)),
+                nn.Conv2d(dim, dim, 3, stride=(2, 2)))
+        elif mode == 'downsample3d':
+            self.resample = nn.Sequential(
+                nn.ZeroPad2d((0, 1, 0, 1)),
+                nn.Conv2d(dim, dim, 3, stride=(2, 2)))
+            self.time_conv = CausalConv3d(
+                dim, dim, (3, 1, 1), stride=(2, 1, 1), padding=(0, 0, 0))
+
+        else:
+            self.resample = nn.Identity()
+
+    def forward(self, x, feat_cache=None, feat_idx=[0]):
+        b, c, t, h, w = x.size()
+        if self.mode == 'upsample3d':
+            if feat_cache is not None:
+                idx = feat_idx[0]
+                if feat_cache[idx] is None:
+                    feat_cache[idx] = 'Rep'
+                    feat_idx[0] += 1
+                else:
+
+                    cache_x = x[:, :, -CACHE_T:, :, :].clone()
+                    if cache_x.shape[2] < 2 and feat_cache[
+                            idx] is not None and feat_cache[idx] != 'Rep':
+                        # cache last frame of last two chunk
+                        cache_x = torch.cat([
+                            feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
+                                cache_x.device), cache_x
+                        ],
+                            dim=2)
+                    if cache_x.shape[2] < 2 and feat_cache[
+                            idx] is not None and feat_cache[idx] == 'Rep':
+                        cache_x = torch.cat([
+                            torch.zeros_like(cache_x).to(cache_x.device),
+                            cache_x
+                        ],
+                            dim=2)
+                    if feat_cache[idx] == 'Rep':
+                        x = self.time_conv(x)
+                    else:
+                        x = self.time_conv(x, feat_cache[idx])
+                    feat_cache[idx] = cache_x
+                    feat_idx[0] += 1
+
+                    x = x.reshape(b, 2, c, t, h, w)
+                    x = torch.stack((x[:, 0, :, :, :, :], x[:, 1, :, :, :, :]),
+                                    3)
+                    x = x.reshape(b, c, t * 2, h, w)
+        t = x.shape[2]
+        x = rearrange(x, 'b c t h w -> (b t) c h w')
+        x = self.resample(x)
+        x = rearrange(x, '(b t) c h w -> b c t h w', t=t)
+
+        if self.mode == 'downsample3d':
+            if feat_cache is not None:
+                idx = feat_idx[0]
+                if feat_cache[idx] is None:
+                    feat_cache[idx] = x.clone()
+                    feat_idx[0] += 1
+                else:
+
+                    cache_x = x[:, :, -1:, :, :].clone()
+                    # if cache_x.shape[2] < 2 and feat_cache[idx] is not None and feat_cache[idx]!='Rep':
+                    #     # cache last frame of last two chunk
+                    #     cache_x = torch.cat([feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(cache_x.device), cache_x], dim=2)
+
+                    x = self.time_conv(
+                        torch.cat([feat_cache[idx][:, :, -1:, :, :], x], 2))
+                    feat_cache[idx] = cache_x
+                    feat_idx[0] += 1
+        return x
+
+    def init_weight(self, conv):
+        conv_weight = conv.weight
+        nn.init.zeros_(conv_weight)
+        c1, c2, t, h, w = conv_weight.size()
+        one_matrix = torch.eye(c1, c2)
+        init_matrix = one_matrix
+        nn.init.zeros_(conv_weight)
+        # conv_weight.data[:,:,-1,1,1] = init_matrix * 0.5
+        conv_weight.data[:, :, 1, 0, 0] = init_matrix  # * 0.5
+        conv.weight.data.copy_(conv_weight)
+        nn.init.zeros_(conv.bias.data)
+
+    def init_weight2(self, conv):
+        conv_weight = conv.weight.data
+        nn.init.zeros_(conv_weight)
+        c1, c2, t, h, w = conv_weight.size()
+        init_matrix = torch.eye(c1 // 2, c2)
+        # init_matrix = repeat(init_matrix, 'o ... -> (o 2) ...').permute(1,0,2).contiguous().reshape(c1,c2)
+        conv_weight[:c1 // 2, :, -1, 0, 0] = init_matrix
+        conv_weight[c1 // 2:, :, -1, 0, 0] = init_matrix
+        conv.weight.data.copy_(conv_weight)
+        nn.init.zeros_(conv.bias.data)
+
+
+class ResidualBlock(nn.Module):
+
+    def __init__(self, in_dim, out_dim, dropout=0.0):
+        super().__init__()
+        self.in_dim = in_dim
+        self.out_dim = out_dim
+
+        # layers
+        self.residual = nn.Sequential(
+            RMS_norm(in_dim, images=False), nn.SiLU(),
+            CausalConv3d(in_dim, out_dim, 3, padding=1),
+            RMS_norm(out_dim, images=False), nn.SiLU(), nn.Dropout(dropout),
+            CausalConv3d(out_dim, out_dim, 3, padding=1))
+        self.shortcut = CausalConv3d(in_dim, out_dim, 1) \
+            if in_dim != out_dim else nn.Identity()
+
+    def forward(self, x, feat_cache=None, feat_idx=[0]):
+        h = self.shortcut(x)
+        for layer in self.residual:
+            if isinstance(layer, CausalConv3d) and feat_cache is not None:
+                idx = feat_idx[0]
+                cache_x = x[:, :, -CACHE_T:, :, :].clone()
+                if cache_x.shape[2] < 2 and feat_cache[idx] is not None:
+                    # cache last frame of last two chunk
+                    cache_x = torch.cat([
+                        feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
+                            cache_x.device), cache_x
+                    ],
+                        dim=2)
+                x = layer(x, feat_cache[idx])
+                feat_cache[idx] = cache_x
+                feat_idx[0] += 1
+            else:
+                x = layer(x)
+        return x + h
+
+
+class AttentionBlock(nn.Module):
+    """
+    Causal self-attention with a single head.
+    """
+
+    def __init__(self, dim):
+        super().__init__()
+        self.dim = dim
+
+        # layers
+        self.norm = RMS_norm(dim)
+        self.to_qkv = nn.Conv2d(dim, dim * 3, 1)
+        self.proj = nn.Conv2d(dim, dim, 1)
+
+        # zero out the last layer params
+        nn.init.zeros_(self.proj.weight)
+
+    def forward(self, x):
+        identity = x
+        b, c, t, h, w = x.size()
+        x = rearrange(x, 'b c t h w -> (b t) c h w')
+        x = self.norm(x)
+        # compute query, key, value
+        q, k, v = self.to_qkv(x).reshape(b * t, 1, c * 3,
+                                         -1).permute(0, 1, 3,
+                                                     2).contiguous().chunk(
+                                                         3, dim=-1)
+
+        # apply attention
+        x = F.scaled_dot_product_attention(
+            q,
+            k,
+            v,
+        )
+        x = x.squeeze(1).permute(0, 2, 1).reshape(b * t, c, h, w)
+
+        # output
+        x = self.proj(x)
+        x = rearrange(x, '(b t) c h w-> b c t h w', t=t)
+        return x + identity
+
+
+class Encoder3d(nn.Module):
+
+    def __init__(self,
+                 dim=128,
+                 z_dim=4,
+                 dim_mult=[1, 2, 4, 4],
+                 num_res_blocks=2,
+                 attn_scales=[],
+                 temperal_downsample=[True, True, False],
+                 dropout=0.0):
+        super().__init__()
+        self.dim = dim
+        self.z_dim = z_dim
+        self.dim_mult = dim_mult
+        self.num_res_blocks = num_res_blocks
+        self.attn_scales = attn_scales
+        self.temperal_downsample = temperal_downsample
+
+        # dimensions
+        dims = [dim * u for u in [1] + dim_mult]
+        scale = 1.0
+
+        # init block
+        self.conv1 = CausalConv3d(3, dims[0], 3, padding=1)
+
+        # downsample blocks
+        downsamples = []
+        for i, (in_dim, out_dim) in enumerate(zip(dims[:-1], dims[1:])):
+            # residual (+attention) blocks
+            for _ in range(num_res_blocks):
+                downsamples.append(ResidualBlock(in_dim, out_dim, dropout))
+                if scale in attn_scales:
+                    downsamples.append(AttentionBlock(out_dim))
+                in_dim = out_dim
+
+            # downsample block
+            if i != len(dim_mult) - 1:
+                mode = 'downsample3d' if temperal_downsample[
+                    i] else 'downsample2d'
+                downsamples.append(Resample(out_dim, mode=mode))
+                scale /= 2.0
+        self.downsamples = nn.Sequential(*downsamples)
+
+        # middle blocks
+        self.middle = nn.Sequential(
+            ResidualBlock(out_dim, out_dim, dropout), AttentionBlock(out_dim),
+            ResidualBlock(out_dim, out_dim, dropout))
+
+        # output blocks
+        self.head = nn.Sequential(
+            RMS_norm(out_dim, images=False), nn.SiLU(),
+            CausalConv3d(out_dim, z_dim, 3, padding=1))
+
+    def forward(self, x, feat_cache=None, feat_idx=[0]):
+        if feat_cache is not None:
+            idx = feat_idx[0]
+            cache_x = x[:, :, -CACHE_T:, :, :].clone()
+            if cache_x.shape[2] < 2 and feat_cache[idx] is not None:
+                # cache last frame of last two chunk
+                cache_x = torch.cat([
+                    feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
+                        cache_x.device), cache_x
+                ],
+                    dim=2)
+            x = self.conv1(x, feat_cache[idx])
+            feat_cache[idx] = cache_x
+            feat_idx[0] += 1
+        else:
+            x = self.conv1(x)
+
+        # downsamples
+        for layer in self.downsamples:
+            if feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)
+            else:
+                x = layer(x)
+
+        # middle
+        for layer in self.middle:
+            if isinstance(layer, ResidualBlock) and feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)
+            else:
+                x = layer(x)
+
+        # head
+        for layer in self.head:
+            if isinstance(layer, CausalConv3d) and feat_cache is not None:
+                idx = feat_idx[0]
+                cache_x = x[:, :, -CACHE_T:, :, :].clone()
+                if cache_x.shape[2] < 2 and feat_cache[idx] is not None:
+                    # cache last frame of last two chunk
+                    cache_x = torch.cat([
+                        feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
+                            cache_x.device), cache_x
+                    ],
+                        dim=2)
+                x = layer(x, feat_cache[idx])
+                feat_cache[idx] = cache_x
+                feat_idx[0] += 1
+            else:
+                x = layer(x)
+        return x
+
+
+class Decoder3d(nn.Module):
+
+    def __init__(self,
+                 dim=128,
+                 z_dim=4,
+                 dim_mult=[1, 2, 4, 4],
+                 num_res_blocks=2,
+                 attn_scales=[],
+                 temperal_upsample=[False, True, True],
+                 dropout=0.0):
+        super().__init__()
+        self.dim = dim
+        self.z_dim = z_dim
+        self.dim_mult = dim_mult
+        self.num_res_blocks = num_res_blocks
+        self.attn_scales = attn_scales
+        self.temperal_upsample = temperal_upsample
+
+        # dimensions
+        dims = [dim * u for u in [dim_mult[-1]] + dim_mult[::-1]]
+        scale = 1.0 / 2**(len(dim_mult) - 2)
+
+        # init block
+        self.conv1 = CausalConv3d(z_dim, dims[0], 3, padding=1)
+
+        # middle blocks
+        self.middle = nn.Sequential(
+            ResidualBlock(dims[0], dims[0], dropout), AttentionBlock(dims[0]),
+            ResidualBlock(dims[0], dims[0], dropout))
+
+        # upsample blocks
+        upsamples = []
+        for i, (in_dim, out_dim) in enumerate(zip(dims[:-1], dims[1:])):
+            # residual (+attention) blocks
+            if i == 1 or i == 2 or i == 3:
+                in_dim = in_dim // 2
+            for _ in range(num_res_blocks + 1):
+                upsamples.append(ResidualBlock(in_dim, out_dim, dropout))
+                if scale in attn_scales:
+                    upsamples.append(AttentionBlock(out_dim))
+                in_dim = out_dim
+
+            # upsample block
+            if i != len(dim_mult) - 1:
+                mode = 'upsample3d' if temperal_upsample[i] else 'upsample2d'
+                upsamples.append(Resample(out_dim, mode=mode))
+                scale *= 2.0
+        self.upsamples = nn.Sequential(*upsamples)
+
+        # output blocks
+        self.head = nn.Sequential(
+            RMS_norm(out_dim, images=False), nn.SiLU(),
+            CausalConv3d(out_dim, 3, 3, padding=1))
+
+    def forward(self, x, feat_cache=None, feat_idx=[0]):
+        # conv1
+        if feat_cache is not None:
+            idx = feat_idx[0]
+            cache_x = x[:, :, -CACHE_T:, :, :].clone()
+            if cache_x.shape[2] < 2 and feat_cache[idx] is not None:
+                # cache last frame of last two chunk
+                cache_x = torch.cat([
+                    feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
+                        cache_x.device), cache_x
+                ],
+                    dim=2)
+            x = self.conv1(x, feat_cache[idx])
+            feat_cache[idx] = cache_x
+            feat_idx[0] += 1
+        else:
+            x = self.conv1(x)
+
+        # middle
+        for layer in self.middle:
+            if isinstance(layer, ResidualBlock) and feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)
+            else:
+                x = layer(x)
+
+        # upsamples
+        for layer in self.upsamples:
+            if feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)
+            else:
+                x = layer(x)
+
+        # head
+        for layer in self.head:
+            if isinstance(layer, CausalConv3d) and feat_cache is not None:
+                idx = feat_idx[0]
+                cache_x = x[:, :, -CACHE_T:, :, :].clone()
+                if cache_x.shape[2] < 2 and feat_cache[idx] is not None:
+                    # cache last frame of last two chunk
+                    cache_x = torch.cat([
+                        feat_cache[idx][:, :, -1, :, :].unsqueeze(2).to(
+                            cache_x.device), cache_x
+                    ],
+                        dim=2)
+                x = layer(x, feat_cache[idx])
+                feat_cache[idx] = cache_x
+                feat_idx[0] += 1
+            else:
+                x = layer(x)
+        return x
+
+
+def count_conv3d(model):
+    count = 0
+    for m in model.modules():
+        if isinstance(m, CausalConv3d):
+            count += 1
+    return count
+
+
+class WanVAE_(nn.Module):
+
+    def __init__(self,
+                 dim=128,
+                 z_dim=4,
+                 dim_mult=[1, 2, 4, 4],
+                 num_res_blocks=2,
+                 attn_scales=[],
+                 temperal_downsample=[True, True, False],
+                 dropout=0.0):
+        super().__init__()
+        self.dim = dim
+        self.z_dim = z_dim
+        self.dim_mult = dim_mult
+        self.num_res_blocks = num_res_blocks
+        self.attn_scales = attn_scales
+        self.temperal_downsample = temperal_downsample
+        self.temperal_upsample = temperal_downsample[::-1]
+
+        # modules
+        self.encoder = Encoder3d(dim, z_dim * 2, dim_mult, num_res_blocks,
+                                 attn_scales, self.temperal_downsample, dropout)
+        self.conv1 = CausalConv3d(z_dim * 2, z_dim * 2, 1)
+        self.conv2 = CausalConv3d(z_dim, z_dim, 1)
+        self.decoder = Decoder3d(dim, z_dim, dim_mult, num_res_blocks,
+                                 attn_scales, self.temperal_upsample, dropout)
+        self.clear_cache()
+
+    def forward(self, x):
+        mu, log_var = self.encode(x)
+        z = self.reparameterize(mu, log_var)
+        x_recon = self.decode(z)
+        return x_recon, mu, log_var
+
+    def encode(self, x, scale):
+        self.clear_cache()
+        # cache
+        t = x.shape[2]
+        iter_ = 1 + (t - 1) // 4
+        # 对encode输入的x，按时间拆分为1、4、4、4....
+        for i in range(iter_):
+            self._enc_conv_idx = [0]
+            if i == 0:
+                out = self.encoder(
+                    x[:, :, :1, :, :],
+                    feat_cache=self._enc_feat_map,
+                    feat_idx=self._enc_conv_idx)
+            else:
+                out_ = self.encoder(
+                    x[:, :, 1 + 4 * (i - 1):1 + 4 * i, :, :],
+                    feat_cache=self._enc_feat_map,
+                    feat_idx=self._enc_conv_idx)
+                out = torch.cat([out, out_], 2)
+        mu, log_var = self.conv1(out).chunk(2, dim=1)
+        if isinstance(scale[0], torch.Tensor):
+            mu = (mu - scale[0].view(1, self.z_dim, 1, 1, 1)) * scale[1].view(
+                1, self.z_dim, 1, 1, 1)
+        else:
+            mu = (mu - scale[0]) * scale[1]
+        self.clear_cache()
+        return mu
+
+    def decode(self, z, scale):
+        self.clear_cache()
+        # z: [b,c,t,h,w]
+        if isinstance(scale[0], torch.Tensor):
+            z = z / scale[1].view(1, self.z_dim, 1, 1, 1) + scale[0].view(
+                1, self.z_dim, 1, 1, 1)
+        else:
+            z = z / scale[1] + scale[0]
+        iter_ = z.shape[2]
+        x = self.conv2(z)
+        for i in range(iter_):
+            self._conv_idx = [0]
+            if i == 0:
+                out = self.decoder(
+                    x[:, :, i:i + 1, :, :],
+                    feat_cache=self._feat_map,
+                    feat_idx=self._conv_idx)
+            else:
+                out_ = self.decoder(
+                    x[:, :, i:i + 1, :, :],
+                    feat_cache=self._feat_map,
+                    feat_idx=self._conv_idx)
+                out = torch.cat([out, out_], 2)
+        self.clear_cache()
+        return out
+
+    def cached_decode(self, z, scale):
+        # z: [b,c,t,h,w]
+        if isinstance(scale[0], torch.Tensor):
+            z = z / scale[1].view(1, self.z_dim, 1, 1, 1) + scale[0].view(
+                1, self.z_dim, 1, 1, 1)
+        else:
+            z = z / scale[1] + scale[0]
+        iter_ = z.shape[2]
+        x = self.conv2(z)
+        for i in range(iter_):
+            self._conv_idx = [0]
+            if i == 0:
+                out = self.decoder(
+                    x[:, :, i:i + 1, :, :],
+                    feat_cache=self._feat_map,
+                    feat_idx=self._conv_idx)
+            else:
+                out_ = self.decoder(
+                    x[:, :, i:i + 1, :, :],
+                    feat_cache=self._feat_map,
+                    feat_idx=self._conv_idx)
+                out = torch.cat([out, out_], 2)
+        return out
+
+    def sample(self, imgs, deterministic=False):
+        mu, log_var = self.encode(imgs)
+        if deterministic:
+            return mu
+        std = torch.exp(0.5 * log_var.clamp(-30.0, 20.0))
+        return mu + std * torch.randn_like(std)
+
+    def clear_cache(self):
+        self._conv_num = count_conv3d(self.decoder)
+        self._conv_idx = [0]
+        self._feat_map = [None] * self._conv_num
+        # cache encode
+        self._enc_conv_num = count_conv3d(self.encoder)
+        self._enc_conv_idx = [0]
+        self._enc_feat_map = [None] * self._enc_conv_num
+
+
+def _video_vae(pretrained_path=None, z_dim=None, device='cpu', **kwargs):
+    """
+    Autoencoder3d adapted from Stable Diffusion 1.x, 2.x and XL.
+    """
+    # params
+    cfg = dict(
+        dim=96,
+        z_dim=z_dim,
+        dim_mult=[1, 2, 4, 4],
+        num_res_blocks=2,
+        attn_scales=[],
+        temperal_downsample=[False, True, True],
+        dropout=0.0)
+    cfg.update(**kwargs)
+
+    # init model
+    with torch.device('meta'):
+        model = WanVAE_(**cfg)
+
+    # load checkpoint
+    logging.info(f'loading {pretrained_path}')
+    model.load_state_dict(
+        torch.load(pretrained_path, map_location=device), assign=True)
+
+    return model
+
+
+class WanVAE:
+
+    def __init__(self,
+                 z_dim=16,
+                 vae_pth='cache/vae_step_411000.pth',
+                 dtype=torch.float,
+                 device="cuda"):
+        self.dtype = dtype
+        self.device = device
+
+        mean = [
+            -0.7571, -0.7089, -0.9113, 0.1075, -0.1745, 0.9653, -0.1517, 1.5508,
+            0.4134, -0.0715, 0.5517, -0.3632, -0.1922, -0.9497, 0.2503, -0.2921
+        ]
+        std = [
+            2.8184, 1.4541, 2.3275, 2.6558, 1.2196, 1.7708, 2.6052, 2.0743,
+            3.2687, 2.1526, 2.8652, 1.5579, 1.6382, 1.1253, 2.8251, 1.9160
+        ]
+        self.mean = torch.tensor(mean, dtype=dtype, device=device)
+        self.std = torch.tensor(std, dtype=dtype, device=device)
+        self.scale = [self.mean, 1.0 / self.std]
+
+        # init model
+        self.model = _video_vae(
+            pretrained_path=vae_pth,
+            z_dim=z_dim,
+        ).eval().requires_grad_(False).to(device)
+
+    def encode(self, videos):
+        """
+        videos: A list of videos each with shape [C, T, H, W].
+        """
+        with amp.autocast(dtype=self.dtype):
+            return [
+                self.model.encode(u.unsqueeze(0), self.scale).float().squeeze(0)
+                for u in videos
+            ]
+
+    def decode(self, zs):
+        with amp.autocast(dtype=self.dtype):
+            return [
+                self.model.decode(u.unsqueeze(0),
+                                  self.scale).float().clamp_(-1, 1).squeeze(0)
+                for u in zs
+            ]