refactor sampling pipeline

Co-authored-by: William Lin <SolitaryThinker@users.noreply.github.com>

.buildkite/pipeline.yml

@@ -61,7 +61,7 @@ steps:
                 - "pyproject.toml"
                 - "docker/Dockerfile.python3.12"
               config:
-                command: "timeout 60m .buildkite/scripts/pr_test.sh"
+                command: "timeout 90m .buildkite/scripts/pr_test.sh"
                 label: "SSIM Tests"
                 env:
                   - TEST_TYPE=ssim

.github/workflows/fastvideo-kernel-publish.yml

@@ -156,8 +156,23 @@ jobs:
           
           # Fix the wheel to be manylinux compliant
           pip install auditwheel
+          # Point auditwheel at torch libs, but do not vendor them into the wheel.
+          TORCH_LIB_DIR=$(python - <<'PY'
+          import os
+          import torch
+
+          print(os.path.join(os.path.dirname(torch.__file__), "lib"))
+          PY
+          )
+          export LD_LIBRARY_PATH="${TORCH_LIB_DIR}:${LD_LIBRARY_PATH}"
           # Target manylinux_2_35 (Ubuntu 22.04 native)
-          auditwheel repair dist/*.whl --plat manylinux_2_35_x86_64 -w fixed_dist
+          auditwheel repair dist/*.whl --plat manylinux_2_35_x86_64 -w fixed_dist \
+            --exclude libtorch_cuda.so \
+            --exclude libtorch_cpu.so \
+            --exclude libtorch.so \
+            --exclude libc10.so \
+            --exclude libc10_cuda.so \
+            --exclude libtorch_python.so
           # Move fixed wheels back to dist for upload consistency
           rm dist/*.whl
           mv fixed_dist/*.whl dist/

.pre-commit-config.yaml

@@ -68,7 +68,7 @@ repos:
     entry: bash
     args:
       - -c
-      - 'git ls-files | grep -v "^fastvideo/tests/ssim/" | grep -v "^fastvideo/tests/inference/lora/L40S_reference_videos/" | grep " " && echo "Filenames should not contain spaces!" && exit 1 || exit 0'
+      - 'git ls-files | grep -v "^\"*fastvideo/tests/ssim/" | grep -v "^\"*fastvideo/tests/inference/lora/L40S_reference_videos/" | grep " " && echo "Filenames should not contain spaces!" && exit 1 || exit 0'
     language: system
     always_run: true
     pass_filenames: false

README.md

@@ -3,7 +3,7 @@
 </div>
 
 <p align="center">
-     | <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/"><b> Quick Start</b></a> | <a href="https://github.com/hao-ai-lab/FastVideo/discussions/982"  target="_blank"><b>Weekly Dev Meeting</b></a>  | 🟣💬 <a href="https://join.slack.com/t/fastvideo/shared_invite/zt-3f4lao1uq-u~Ipx6Lt4J27AlD2y~IdLQ" target="_blank"> <b>Slack</b> </a> |  🟣💬 <a href="https://ibb.co/c7g1qdD" target="_blank"> <b> WeChat </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/"><b> Quick Start</b></a> | <a href="https://github.com/hao-ai-lab/FastVideo/discussions/982"  target="_blank"><b>Weekly Dev Meeting</b></a>  | 🟣💬 <a href="https://join.slack.com/t/fastvideo/shared_invite/zt-3f4lao1uq-u~Ipx6Lt4J27AlD2y~IdLQ" target="_blank"> <b>Slack</b> </a> |  🟣💬 <a href="https://ibb.co/sv3MMKyv" target="_blank"> <b> WeChat </b> </a> |
 </p>
 
 **FastVideo is a unified post-training and inference framework for accelerated video generation.**

docs/attention/index.md

@@ -41,7 +41,7 @@ Clone the repository and build the kernel:
 
 ```bash
 # Clone recursively to get ThunderKittens submodule
-git clone --recursive https://github.com/hao-ai-lab/FastVideo.git
+git clone https://github.com/hao-ai-lab/FastVideo.git
 cd FastVideo/fastvideo-kernel
 
 # Build and install

docs/attention/vsa/index.md

@@ -13,11 +13,13 @@ from fastvideo_kernel import video_sparse_attn
 
 # q, k, v: [batch_size, num_heads, seq_len, head_dim]
 # variable_block_sizes: Number of valid tokens per block
+# q_variable_block_sizes: Number of valid tokens per q block (can differ from KV for q/k of different lengths)
 # topk: Number of blocks to attend
 
 output = video_sparse_attn(
     q, k, v, 
-    variable_block_sizes=block_sizes, 
+    block_sizes,
+    block_sizes,
     topk=32
 )
 ```

docs/getting_started/installation.md

@@ -11,15 +11,13 @@ FastVideo supports the following hardware platforms:
 ### Using pip
 
 ```bash
+# Create and activate a new conda environment
+conda create -n fastvideo python=3.12
+conda activate fastvideo
+
 pip install fastvideo
 ```
 
-### Using conda
-
-```bash
-conda install -c conda-forge fastvideo
-```
-
 ### From source
 
 ```bash
@@ -28,6 +26,12 @@ cd FastVideo
 pip install -e .
 ```
 
+Also optionally install flash-attn:
+
+```bash
+pip install flash-attn --no-build-isolation
+```
+
 ## Hardware Requirements
 
 - **NVIDIA GPUs**: CUDA 11.8+ with compute capability 7.0+

docs/getting_started/quick_start.md

@@ -15,6 +15,12 @@ conda activate fastvideo
 pip install fastvideo
 ```
 
+Also optionally install flash-attn:
+
+```bash
+pip install flash-attn --no-build-isolation
+```
+
 ## Basic Usage
 
 ### Text-to-Video Generation

docs/inference/inference_quick_start.md

@@ -106,6 +106,8 @@ If you encounter CUDA out of memory errors:
 - Enable memory optimization with `enable_model_cpu_offload`
 - Try a smaller model or use distilled versions
 - Use `num_gpus` > 1 if multiple GPUs are available
+- Try enabling FSDP inference with `use_fsdp_inference=True` (may slow down generation)
+- Try enabling DiT layerwise offload with `dit_layerwise_offload=True` (now only a few models support this, but may introduce less overhead than FSDP)
 
 ### Slow Generation
 

docs/utilities/lora.md

@@ -0,0 +1,71 @@
+# LoRA Extraction and Merging
+
+Tools for extracting and merging LoRA adapters for FastVideo models.
+
+## Extract LoRA Adapter
+
+```bash
+python scripts/lora_extraction/extract_lora.py \
+  --base Wan-AI/Wan2.2-TI2V-5B-Diffusers \
+  --ft FastVideo/FastWan2.2-TI2V-5B-FullAttn-Diffusers \
+  --out adapter_r32.safetensors \
+  --rank 32
+```
+
+**Options:**
+
+- `--base`: Base model (HuggingFace ID or local path)
+- `--ft`: Fine-tuned model (HuggingFace ID or local path)
+- `--out`: Output adapter file
+- `--rank`: LoRA rank (16, 32, 64, 128)
+- `--full-rank`: Extract full-rank adapter (optional)
+
+## Merge Adapter
+
+```bash
+python scripts/lora_extraction/merge_lora.py \
+  --base Wan-AI/Wan2.2-TI2V-5B-Diffusers \
+  --adapter adapter_r32.safetensors \
+  --ft FastVideo/FastWan2.2-TI2V-5B-FullAttn-Diffusers \
+  --output merged_model
+```
+
+**Options:**
+
+- `--base`: Base model (HuggingFace ID or local path)
+- `--adapter`: LoRA adapter file (.safetensors)
+- `--ft`: Fine-tuned model (for configuration)
+- `--output`: Output directory
+
+## Validate Quality (Optional)
+
+```bash
+python scripts/lora_extraction/lora_inference_comparison.py \
+  --base merged_model \
+  --ft FastVideo/FastWan2.2-TI2V-5B-FullAttn-Diffusers \
+  --adapter NONE \
+  --output-dir results \
+  --prompt "A cat sitting on a windowsill" \
+  --seed 42 \
+  --height 480 \
+  --width 480 \
+  --num-frames 49 \
+  --num-inference-steps 32 \
+  --compute-ssim \
+  --compute-lpips
+```
+
+**Options:**
+
+- `--base`: Merged model or base model path
+- `--ft`: Fine-tuned model (reference)
+- `--adapter`: Path to adapter or NONE
+- `--output-dir`: Output directory
+- `--prompt`: Text prompt (default: "A cat sitting on a windowsill")
+- `--seed`: Random seed (default: 42)
+- `--height`: Video height (default: 480)
+- `--width`: Video width (default: 832)
+- `--num-frames`: Number of frames (default: 49)
+- `--num-inference-steps`: Inference steps (default: 32)
+- `--compute-ssim`: Compute SSIM metric
+- `--compute-lpips`: Compute LPIPS metric

examples/inference/basic/basic.py

@@ -12,7 +12,7 @@ def main():
         "Wan-AI/Wan2.1-T2V-1.3B-Diffusers",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=False,
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_cosmos2_5_t2w.py

@@ -0,0 +1,42 @@
+from fastvideo import VideoGenerator
+
+
+def main():
+    # Point this to your local diffusers model dir (or replace with a HF model ID).
+    model_path = "KyleShao/Cosmos-Predict2.5-2B-Diffusers"
+
+    generator = VideoGenerator.from_pretrained(
+        model_path,
+        num_gpus=1,
+        use_fsdp_inference=False,  # set True if GPU is out of memory
+        dit_cpu_offload=False,
+        vae_cpu_offload=False,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=True,
+    )
+
+    prompt = (
+       "A high-definition video captures the precision of robotic welding in an industrial setting. The first frame showcases a robotic arm, equipped with a welding torch, positioned over a large metal structure. The welding process is in full swing, with bright sparks and intense light illuminating the scene, creating a vivid display of blue and white hues. A significant amount of smoke billows around the welding area, partially obscuring the view but emphasizing the heat and activity. The background reveals parts of the workshop environment, including a ventilation system and various pieces of machinery, indicating a busy and functional industrial workspace. As the video progresses, the robotic arm maintains its steady position, continuing the welding process and moving to its left. The welding torch consistently emits sparks and light, and the smoke continues to rise, diffusing slightly as it moves upward. The metal surface beneath the torch shows ongoing signs of heating and melting. The scene retains its industrial ambiance, with the welding sparks and smoke dominating the visual field, underscoring the ongoing nature of the welding operation."
+    )
+
+    video = generator.generate_video(
+        prompt,
+        negative_prompt="",
+        height=704,
+        width=1280,
+        num_frames=77,
+        num_inference_steps=35,
+        guidance_scale=7.0,
+        fps=24,
+        output_path="outputs_video/cosmos2_5_t2w.mp4",
+        save_video=True,
+    )
+
+    generator.shutdown()
+
+
+if __name__ == "__main__":
+    main()
+
+
+

examples/inference/basic/basic_dmd.py

@@ -14,7 +14,7 @@ def main():
         model_name,
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         # Adjust these offload parameters if you have < 32GB of VRAM
         text_encoder_cpu_offload=True,
         pin_cpu_memory=True, # set to false if low CPU RAM or hit obscure "CUDA error: Invalid argument"

examples/inference/basic/basic_hy15.py

@@ -12,7 +12,7 @@ def main():
         "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True,
         vae_cpu_offload=True,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_longcat_i2v.py

@@ -0,0 +1,210 @@
+"""
+LongCat Image-to-Video (I2V) Example Script
+
+This script demonstrates LongCat I2V inference using the FastVideo Python API.
+LongCat I2V takes an input image and generates a video from it.
+
+It runs both basic generation (50 steps) and distill+refine generation
+(16 steps distill + 50 steps refinement to 720p with BSA).
+
+Usage:
+    python examples/inference/basic/basic_longcat_i2v.py
+
+Note:
+    Refinement uses 768x768 dimensions where latent (48x48) is divisible by 8,
+    compatible with BSA chunks [4, 4, 8].
+"""
+
+import glob
+import os
+
+from fastvideo import VideoGenerator
+
+# Common prompts and settings matching the shell script examples
+PROMPT = (
+    "A woman sits at a wooden table by the window in a cozy café. She reaches out "
+    "with her right hand, picks up the white coffee cup from the saucer, and gently "
+    "brings it to her lips to take a sip. After drinking, she places the cup back on "
+    "the table and looks out the window, enjoying the peaceful atmosphere."
+)
+
+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"
+)
+
+# Input image path
+IMAGE_PATH = "assets/girl.png"
+
+SEED = 42
+
+
+def basic_generation():
+    """
+    Run basic LongCat I2V generation (50 steps at 480p).
+    
+    This uses the full 50-step denoising process for highest quality.
+    """
+    print("=" * 60)
+    print("LongCat I2V: Basic Generation (50 steps, 480p)")
+    print("=" * 60)
+    
+    generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-I2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=False,
+    )
+    
+    output_path = "outputs_video/longcat_i2v_basic"
+    
+    generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        image_path=IMAGE_PATH,
+        output_path=output_path,
+        save_video=True,
+        height=480,
+        width=480,  # Square
+        num_frames=93,
+        num_inference_steps=50,
+        fps=15,
+        guidance_scale=4.0,
+        seed=SEED,
+    )
+    
+    print(f"\nBasic generation complete! Video saved to: {output_path}")
+    generator.shutdown()
+
+
+def distill_refine_generation():
+    """
+    Run LongCat I2V with distill+refine pipeline (16 steps + refinement to 768p).
+    
+    This uses the distilled LoRA for fast 480p generation (16 steps),
+    then refines to 768p using the refinement LoRA with BSA enabled.
+    """
+    print("\n" + "=" * 60)
+    print("LongCat I2V: Distill + Refine Pipeline")
+    print("=" * 60)
+    
+    # Stage 1: Distilled generation (16 steps at 480p)
+    print("\n[Stage 1] Distilled generation (16 steps, 480p)")
+    print("-" * 40)
+    
+    generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-I2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=True,
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=False,
+        lora_path="FastVideo/LongCat-Video-T2V-Distilled-LoRA",
+        lora_nickname="distilled",
+    )
+    
+    distill_output_path = "outputs_video/longcat_i2v_distill"
+    
+    generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        image_path=IMAGE_PATH,
+        output_path=distill_output_path,
+        save_video=True,
+        height=480,
+        width=480,  # Square
+        num_frames=93,
+        num_inference_steps=16,
+        fps=15,
+        guidance_scale=1.0,
+        seed=SEED,
+    )
+    
+    print(f"Distilled generation complete! Video saved to: {distill_output_path}")
+    generator.shutdown()
+    
+    # Stage 2: Refinement (480p -> 768p)
+    print("\n[Stage 2] Refinement (480p -> 768p with BSA)")
+    print("-" * 40)
+    
+    # Find the actual saved video file from stage 1
+    video_files = glob.glob(os.path.join(distill_output_path, "*.mp4"))
+    if not video_files:
+        raise FileNotFoundError(f"No video file found in {distill_output_path}")
+    # Use the most recently created video file
+    distill_video_path = max(video_files, key=os.path.getmtime)
+    print(f"Using stage 1 video: {distill_video_path}")
+    
+    # Create a new generator with refinement LoRA and BSA enabled
+    # Note: Refinement uses the T2V model (not I2V) since it's upscaling the generated video
+    # For BSA [4, 4, 8]: latent must be divisible by 8
+    # 768x768: latent 48x48, 48%8=0 ✓
+    refine_generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-T2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=True,
+        dit_cpu_offload=True,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=True,
+        bsa_sparsity=0.875,
+        bsa_chunk_q=[4, 4, 4],
+        bsa_chunk_k=[4, 4, 4],
+        lora_path="FastVideo/LongCat-Video-T2V-Refinement-LoRA",
+        lora_nickname="refinement",
+    )
+    
+    refine_output_path = "outputs_video/longcat_i2v_refine_720p"
+    
+    refine_generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        output_path=refine_output_path,
+        save_video=True,
+        refine_from=distill_video_path,
+        t_thresh=0.5,
+        spatial_refine_only=False,
+        num_cond_frames=0,
+        height=720,
+        width=720,
+        num_inference_steps=50,
+        fps=30,
+        guidance_scale=1.0,
+        seed=SEED,
+    )
+    
+    print(f"Refinement complete! Video saved to: {refine_output_path}")
+    refine_generator.shutdown()
+
+
+def main():
+    """Run both basic and distill+refine generation pipelines."""
+    print("\n" + "=" * 60)
+    print("LongCat Image-to-Video Example")
+    print("=" * 60 + "\n")
+    
+    # Run basic generation
+    basic_generation()
+    
+    # Run distill+refine pipeline
+    distill_refine_generation()
+    
+    print("\n" + "=" * 60)
+    print("All generations complete!")
+    print("=" * 60)
+
+
+if __name__ == "__main__":
+    main()
+
+

examples/inference/basic/basic_longcat_t2v.py

@@ -0,0 +1,198 @@
+"""
+LongCat Text-to-Video (T2V) Example Script
+
+This script demonstrates LongCat T2V inference using the FastVideo Python API.
+It runs both basic generation (50 steps) and distill+refine generation
+(16 steps distill + 50 steps refinement to 720p).
+
+Usage:
+    python examples/inference/basic/basic_longcat_t2v.py
+"""
+
+import glob
+import os
+
+from fastvideo import VideoGenerator
+
+# Common prompts and settings matching the shell script examples
+PROMPT = (
+    "In a realistic photography style, a white boy around seven or eight years old "
+    "sits on a park bench, wearing a light blue T-shirt, denim shorts, and white sneakers. "
+    "He holds an ice cream cone with vanilla and chocolate flavors, and beside him is a "
+    "medium-sized golden Labrador. Smiling, the boy offers the ice cream to the dog, "
+    "who eagerly licks it with its tongue. The sun is shining brightly, and the background "
+    "features a green lawn and several tall trees, creating a warm and loving scene."
+)
+
+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 = 42
+
+
+def basic_generation():
+    """
+    Run basic LongCat T2V generation (50 steps at 480p).
+    
+    This uses the full 50-step denoising process for highest quality.
+    """
+    print("=" * 60)
+    print("LongCat T2V: Basic Generation (50 steps, 480p)")
+    print("=" * 60)
+    
+    generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-T2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=False,
+    )
+    
+    output_path = "outputs_video/longcat_t2v_basic"
+    
+    generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        output_path=output_path,
+        save_video=True,
+        height=480,
+        width=832,
+        num_frames=93,
+        num_inference_steps=50,
+        fps=15,
+        guidance_scale=4.0,
+        seed=SEED,
+    )
+    
+    print(f"\nBasic generation complete! Video saved to: {output_path}")
+    generator.shutdown()
+
+
+def distill_refine_generation():
+    """
+    Run LongCat T2V with distill+refine pipeline (16 steps + refinement to 720p).
+    
+    This uses the distilled LoRA for fast 480p generation (16 steps),
+    then refines to 720p using the refinement LoRA with BSA enabled.
+    """
+    print("\n" + "=" * 60)
+    print("LongCat T2V: Distill + Refine Pipeline")
+    print("=" * 60)
+    
+    # Stage 1: Distilled generation (16 steps at 480p)
+    print("\n[Stage 1] Distilled generation (16 steps, 480p)")
+    print("-" * 40)
+    
+    generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-T2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=True,
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=False,
+        lora_path="FastVideo/LongCat-Video-T2V-Distilled-LoRA",
+        lora_nickname="distilled",
+    )
+    
+    distill_output_path = "outputs_video/longcat_t2v_distill"
+    
+    generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        output_path=distill_output_path,
+        save_video=True,
+        height=480,
+        width=832,
+        num_frames=93,
+        num_inference_steps=16,
+        fps=15,
+        guidance_scale=1.0,
+        seed=SEED,
+    )
+    
+    print(f"Distilled generation complete! Video saved to: {distill_output_path}")
+    generator.shutdown()
+    
+    # Stage 2: Refinement (480p -> 720p)
+    print("\n[Stage 2] Refinement (480p -> 720p with BSA)")
+    print("-" * 40)
+    
+    # Find the actual saved video file from stage 1
+    video_files = glob.glob(os.path.join(distill_output_path, "*.mp4"))
+    if not video_files:
+        raise FileNotFoundError(f"No video file found in {distill_output_path}")
+    # Use the most recently created video file
+    distill_video_path = max(video_files, key=os.path.getmtime)
+    print(f"Using stage 1 video: {distill_video_path}")
+    
+    # Create a new generator with refinement LoRA and BSA enabled
+    refine_generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-T2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=True,
+        dit_cpu_offload=True,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=True,
+        bsa_sparsity=0.875,
+        bsa_chunk_q=[4, 4, 8],
+        bsa_chunk_k=[4, 4, 8],
+        lora_path="FastVideo/LongCat-Video-T2V-Refinement-LoRA",
+        lora_nickname="refinement",
+    )
+    
+    refine_output_path = "outputs_video/longcat_t2v_refine_720p"
+    
+    refine_generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        output_path=refine_output_path,
+        save_video=True,
+        refine_from=distill_video_path,
+        t_thresh=0.5,
+        spatial_refine_only=False,
+        num_cond_frames=0,
+        height=720,
+        width=1280,
+        num_inference_steps=50,
+        fps=30,
+        guidance_scale=1.0,
+        seed=SEED,
+    )
+    
+    print(f"Refinement complete! Video saved to: {refine_output_path}")
+    refine_generator.shutdown()
+
+
+def main():
+    """Run both basic and distill+refine generation pipelines."""
+    print("\n" + "=" * 60)
+    print("LongCat Text-to-Video Example")
+    print("=" * 60 + "\n")
+    
+    # Run basic generation
+    basic_generation()
+    
+    # Run distill+refine pipeline
+    distill_refine_generation()
+    
+    print("\n" + "=" * 60)
+    print("All generations complete!")
+    print("=" * 60)
+
+
+if __name__ == "__main__":
+    main()
+
+

examples/inference/basic/basic_longcat_vc.py

@@ -0,0 +1,228 @@
+"""
+LongCat Video Continuation (VC) Example Script
+
+This script demonstrates LongCat VC inference using the FastVideo Python API.
+LongCat VC takes an input video and generates a continuation of it.
+
+It runs both basic generation (50 steps) and distill+refine generation
+(16 steps distill + 50 steps refinement to 720p).
+
+Usage:
+    python examples/inference/basic/basic_longcat_vc.py
+
+Prerequisites:
+    - Ensure the input video exists at assets/motorcycle.mp4
+      (or provide your own video)
+"""
+
+import glob
+import os
+
+from fastvideo import VideoGenerator
+
+# Common prompts and settings matching the shell script examples
+PROMPT = (
+    "A person rides a motorcycle along a long, straight road that stretches between "
+    "a body of water and a forested hillside. The rider steadily accelerates, keeping "
+    "the motorcycle centered between the guardrails, while the scenery passes by on "
+    "both sides. The video captures the journey from the rider's perspective, emphasizing "
+    "the sense of motion and adventure."
+)
+
+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"
+)
+
+# Input video path
+VIDEO_PATH = "assets/motorcycle.mp4"
+
+# Number of conditioning frames from the input video
+NUM_COND_FRAMES = 13
+
+SEED = 42
+
+
+def basic_generation():
+    """
+    Run basic LongCat VC generation (50 steps at 480p).
+    
+    This uses the full 50-step denoising process for highest quality.
+    """
+    print("=" * 60)
+    print("LongCat VC: Basic Generation (50 steps, 480p)")
+    print("=" * 60)
+    
+    # Check if video exists
+    if not os.path.exists(VIDEO_PATH):
+        raise FileNotFoundError(
+            f"Video not found at {VIDEO_PATH}. "
+            "Please provide a valid video path."
+        )
+    
+    generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-VC-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=False,
+    )
+    
+    output_path = "outputs_video/longcat_vc_basic"
+    
+    generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        video_path=VIDEO_PATH,
+        num_cond_frames=NUM_COND_FRAMES,
+        output_path=output_path,
+        save_video=True,
+        height=480,
+        width=832,
+        num_frames=93,
+        num_inference_steps=50,
+        fps=15,
+        guidance_scale=4.0,
+        seed=SEED,
+    )
+    
+    print(f"\nBasic generation complete! Video saved to: {output_path}")
+    generator.shutdown()
+
+
+def distill_refine_generation():
+    """
+    Run LongCat VC with distill+refine pipeline (16 steps + refinement to 720p).
+    
+    This uses the distilled LoRA for fast 480p generation (16 steps),
+    then refines to 720p using the refinement LoRA with BSA enabled.
+    """
+    print("\n" + "=" * 60)
+    print("LongCat VC: Distill + Refine Pipeline")
+    print("=" * 60)
+    
+    # Check if video exists
+    if not os.path.exists(VIDEO_PATH):
+        raise FileNotFoundError(
+            f"Video not found at {VIDEO_PATH}. "
+            "Please provide a valid video path."
+        )
+    
+    # Stage 1: Distilled generation (16 steps at 480p)
+    print("\n[Stage 1] Distilled generation (16 steps, 480p)")
+    print("-" * 40)
+    
+    generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-VC-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=True,
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=False,
+        lora_path="FastVideo/LongCat-Video-T2V-Distilled-LoRA",
+        lora_nickname="distilled",
+    )
+    
+    distill_output_path = "outputs_video/longcat_vc_distill"
+    
+    generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        video_path=VIDEO_PATH,
+        num_cond_frames=NUM_COND_FRAMES,
+        output_path=distill_output_path,
+        save_video=True,
+        height=480,
+        width=832,
+        num_frames=93,
+        num_inference_steps=16,
+        fps=15,
+        guidance_scale=1.0,
+        seed=SEED,
+    )
+    
+    print(f"Distilled generation complete! Video saved to: {distill_output_path}")
+    generator.shutdown()
+    
+    # Stage 2: Refinement (480p -> 720p)
+    print("\n[Stage 2] Refinement (480p -> 720p with BSA)")
+    print("-" * 40)
+    
+    # Find the actual saved video file from stage 1
+    video_files = glob.glob(os.path.join(distill_output_path, "*.mp4"))
+    if not video_files:
+        raise FileNotFoundError(f"No video file found in {distill_output_path}")
+    # Use the most recently created video file
+    distill_video_path = max(video_files, key=os.path.getmtime)
+    print(f"Using stage 1 video: {distill_video_path}")
+    
+    # Create a new generator with refinement LoRA and BSA enabled
+    # Note: Refinement uses the T2V model (not VC) since it's upscaling the generated video
+    refine_generator = VideoGenerator.from_pretrained(
+        "FastVideo/LongCat-Video-T2V-Diffusers",
+        num_gpus=1,
+        use_fsdp_inference=True,
+        dit_cpu_offload=True,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=False,
+        enable_bsa=True,
+        bsa_sparsity=0.875,
+        bsa_chunk_q=[4, 4, 8],
+        bsa_chunk_k=[4, 4, 8],
+        lora_path="FastVideo/LongCat-Video-T2V-Refinement-LoRA",
+        lora_nickname="refinement",
+    )
+    
+    refine_output_path = "outputs_video/longcat_vc_refine_720p"
+    
+    refine_generator.generate_video(
+        prompt=PROMPT,
+        negative_prompt=NEGATIVE_PROMPT,
+        output_path=refine_output_path,
+        save_video=True,
+        refine_from=distill_video_path,
+        t_thresh=0.5,
+        spatial_refine_only=False,
+        num_cond_frames=0,  # For refinement, no conditioning frames
+        height=720,
+        width=1280,
+        num_inference_steps=50,
+        fps=30,
+        guidance_scale=1.0,
+        seed=SEED,
+    )
+    
+    print(f"Refinement complete! Video saved to: {refine_output_path}")
+    refine_generator.shutdown()
+
+
+def main():
+    """Run both basic and distill+refine generation pipelines."""
+    print("\n" + "=" * 60)
+    print("LongCat Video Continuation Example")
+    print("=" * 60 + "\n")
+    
+    # Run basic generation
+    basic_generation()
+    
+    # Run distill+refine pipeline
+    distill_refine_generation()
+    
+    print("\n" + "=" * 60)
+    print("All generations complete!")
+    print("=" * 60)
+
+
+if __name__ == "__main__":
+    main()
+
+

examples/inference/basic/basic_matrixgame.py

@@ -43,7 +43,7 @@ def main():
         config["model_path"],
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_matrixgame_streaming.py

@@ -46,7 +46,7 @@ async def main():
         config["model_path"],
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_self_forcing_causal.py

@@ -13,7 +13,7 @@ def main():
         model_name,
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         text_encoder_cpu_offload=False,
         dit_cpu_offload=False,
     )

examples/inference/basic/basic_self_forcing_causal_wan2_2_i2v.py

@@ -14,7 +14,7 @@ def main():
         "FastVideo/SFWan2.2-I2V-A14B-Preview-Diffusers",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         dit_precision="fp32",
         vae_cpu_offload=False,

examples/inference/basic/basic_self_forcing_causal_wan2_2_t2v.py

@@ -14,7 +14,7 @@ def main():
         "rand0nmr/SFWan2.2-T2V-A14B-Diffusers",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_turbodiffusion.py

@@ -15,12 +15,14 @@ def main() -> None:
         "loayrashid/TurboWan2.1-T2V-1.3B-Diffusers",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        # TurboDiffusion uses a custom pipeline with RCM scheduler
-        override_pipeline_cls_name="TurboDiffusionPipeline",
+        use_fsdp_inference=False, # set to True if GPU is out of memory
+
+        # set to false if using RTX 4090 
+        # pin_cpu_memory=False,
     )
 
     # Generate videos with the same simple API, regardless of GPU count
-    # TurboDiffusion uses guidance_scale=1.0 (no CFG) and only 4 steps
+    # TurboDiffusion defaults: guidance_scale=1.0 and num_inference_steps=4 (from config)
     prompt = (
         "A curious raccoon peers through a vibrant field of yellow sunflowers, its eyes "
         "wide with interest. The playful yet serene atmosphere is complemented by soft "
@@ -30,9 +32,7 @@ def main() -> None:
         prompt,
         output_path=OUTPUT_PATH,
         save_video=True,
-        num_inference_steps=4,
         seed=42,
-        guidance_scale=1.0,
     )
 
     # Generate another video with a different prompt, without reloading the model!
@@ -47,9 +47,7 @@ def main() -> None:
         prompt2,
         output_path=OUTPUT_PATH,
         save_video=True,
-        num_inference_steps=4,
         seed=42,
-        guidance_scale=1.0,
     )
 
 

examples/inference/basic/basic_turbodiffusion_14b.py

@@ -15,8 +15,6 @@ def main() -> None:
         "loayrashid/TurboWan2.1-T2V-14B-Diffusers",
         # 14B model needs more GPUs
         num_gpus=2,
-        # TurboDiffusion uses a custom pipeline with RCM scheduler
-        override_pipeline_cls_name="TurboDiffusionPipeline",
     )
 
     prompt = (
@@ -28,9 +26,7 @@ def main() -> None:
         prompt,
         output_path=OUTPUT_PATH,
         save_video=True,
-        num_inference_steps=4,
         seed=42,
-        guidance_scale=1.0,
     )
 
     # Generate another video with a different prompt, without reloading the model!
@@ -45,9 +41,7 @@ def main() -> None:
         prompt2,
         output_path=OUTPUT_PATH,
         save_video=True,
-        num_inference_steps=4,
         seed=42,
-        guidance_scale=1.0,
     )
 
 

examples/inference/basic/basic_turbodiffusion_i2v.py

@@ -0,0 +1,37 @@
+import os
+
+# Set SLA attention backend BEFORE fastvideo imports
+os.environ["FASTVIDEO_ATTENTION_BACKEND"] = "SLA_ATTN"
+
+from fastvideo import VideoGenerator
+
+# Use local model path
+MODEL_PATH = "loayrashid/TurboWan2.2-I2V-A14B-Diffusers"
+OUTPUT_PATH = "video_samples_turbodiffusion_i2v"
+
+
+def main() -> None:
+    # TurboDiffusion I2V: 1-4 step image-to-video generation
+    generator = VideoGenerator.from_pretrained(
+        MODEL_PATH,
+        num_gpus=2,
+    )
+
+    # Example prompt and image for I2V
+    prompt = ("Summer beach vacation style, a white cat wearing sunglasses sits on a surfboard. The fluffy-furred feline gazes directly at the camera with a relaxed expression. Blurred beach scenery forms the background featuring crystal-clear waters, distant green hills, and a blue sky dotted with white clouds. The cat assumes a naturally relaxed posture, as if savoring the sea breeze and warm sunlight. A close-up shot highlights the feline's intricate details and the refreshing atmosphere of the seaside.")
+
+    # Use an example image path
+    image_path = "https://huggingface.co/datasets/YiYiXu/testing-images/resolve/main/wan_i2v_input.JPG"
+
+
+    video = generator.generate_video(
+        prompt,
+        image_path=image_path,
+        output_path=OUTPUT_PATH,
+        save_video=True,
+        seed=42,
+    )
+
+
+if __name__ == "__main__":
+    main()

examples/inference/basic/basic_wan2_2.py

@@ -12,7 +12,7 @@ def main():
         "Wan-AI/Wan2.2-T2V-A14B-Diffusers",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory 
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_wan2_2_Fun.py

@@ -14,7 +14,7 @@ def main():
         # "alibaba-pai/Wan2.2-Fun-A14B-Control",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_wan2_2_i2v.py

@@ -12,7 +12,7 @@ def main():
         "Wan-AI/Wan2.2-I2V-A14B-Diffusers",
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True, # DiT need to be offloaded for MoE
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/inference/basic/basic_wan2_2_ti2v.py

@@ -11,7 +11,7 @@ def main():
         model_name,
         # FastVideo will automatically handle distributed setup
         num_gpus=1,
-        use_fsdp_inference=True,
+        use_fsdp_inference=False, # set to True if GPU is out of memory
         dit_cpu_offload=True,
         vae_cpu_offload=False,
         text_encoder_cpu_offload=True,

examples/training/rl/finetune_t2v_grpo.sh

@@ -0,0 +1,129 @@
+#!/bin/bash
+
+# Change to FastVideo root directory (3 levels up from this script)
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+FASTVIDEO_ROOT="$(cd "$SCRIPT_DIR/../../.." && pwd)"
+cd "$FASTVIDEO_ROOT"
+
+# Add FastVideo root to PYTHONPATH so Python can find the fastvideo package
+export PYTHONPATH="$FASTVIDEO_ROOT${PYTHONPATH:+:$PYTHONPATH}"
+
+export WANDB_BASE_URL="https://api.wandb.ai"
+export WANDB_MODE=online
+# export FASTVIDEO_ATTENTION_BACKEND=TORCH_SDPA
+
+MODEL_PATH="Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
+RL_DATASET_DIR="data/ocr/"  # Path to RL prompt dataset directory (should contain train.txt and test.txt)
+VALIDATION_DATASET_FILE="$SCRIPT_DIR/validation.json"
+NUM_GPUS=1
+
+# use GPU 3
+export CUDA_VISIBLE_DEVICES=3
+
+
+# Training arguments
+training_args=(
+  --tracker_project_name "wan_t2v_grpo"
+  --output_dir "checkpoints/wan_t2v_grpo"
+  --max_train_steps 5000
+  --train_batch_size 4
+  # --train_sp_batch_size 4
+  --train_sp_batch_size 1
+  --gradient_accumulation_steps 1
+  --num_latent_t 5
+  --num_height 240
+  --num_width 416
+  --num_frames 33
+  --lora_rank 32
+  --lora_training True
+)
+
+# Parallel arguments
+parallel_args=(
+  --num_gpus $NUM_GPUS 
+  --sp_size $NUM_GPUS 
+  --tp_size $NUM_GPUS
+  --hsdp_replicate_dim 1
+  --hsdp_shard_dim $NUM_GPUS
+  # --use-fsdp-inference False
+)
+
+# Model arguments
+model_args=(
+  --model_path $MODEL_PATH
+  --pretrained_model_name_or_path $MODEL_PATH
+)
+
+# Dataset arguments (for RL prompt dataset)
+dataset_args=(
+  --data_path $RL_DATASET_DIR  # Used as fallback if rl_dataset_path not set
+  --rl_dataset_path $RL_DATASET_DIR  # RL prompt dataset directory
+  --rl_dataset_type "text"  # "text" or "geneval"
+  --rl_num_image_per_prompt 4  # k parameter (number of samples per prompt)
+  --dataloader_num_workers 1
+)
+
+# Validation arguments
+validation_args=(
+  --log_validation True
+  --validation_dataset_file $VALIDATION_DATASET_FILE
+  --validation_steps 5
+  --validation_sampling_steps "50" 
+  --validation_guidance_scale "6.0"
+)
+
+# Optimizer arguments
+optimizer_args=(
+  --learning_rate 5e-5
+  --mixed_precision "bf16"
+  --weight_only_checkpointing_steps 10
+  --training_state_checkpointing_steps 10
+  --weight_decay 1e-4
+  --max_grad_norm 1.0
+)
+
+# RL-specific arguments
+rl_args=(
+  --inference_mode False
+  --rl_mode True
+  --rl_algorithm "grpo"
+  --rl_kl_beta 0.004  # KL regularization coefficient
+  --rl_policy_clip_range 0.2  # Policy clipping range for GRPO
+  --rl_kl_reward 0.0  # KL reward coefficient (typically 0)
+  --rl_global_std False  # Use per-prompt std (recommended for GRPO)
+  --rl_per_prompt_stat_tracking True  # Enable per-prompt stat tracking
+  --rl_warmup_steps 0  # Number of warmup steps (SFT before RL)
+  --reward-models "{\"paddle_ocr\": 1.0}" # use video_ocr reward function
+)
+
+# CFG arguments
+cfg_args=(
+  --guidance_scale 1.0 # use guidance_scale > 1.0 to enable CFG
+)
+
+# Miscellaneous arguments
+miscellaneous_args=(
+  --inference_mode False
+  --checkpoints_total_limit 3
+  --training_cfg_rate 0.0  # No CFG during training (CFG used in sampling)
+  --dit_precision "fp32"
+  # --dit_precision "bf16"
+  --num_euler_timesteps 50
+  --ema_start_step 0
+  # --resume_from_checkpoint "checkpoints/wan_t2v_grpo/checkpoint-XXX"
+  --enable-gradient-checkpointing-type "full"
+)
+
+torchrun \
+  --nnodes 1 \
+  --nproc_per_node $NUM_GPUS \
+  --master_port 29501 \
+    "$FASTVIDEO_ROOT/fastvideo/training/wan_rl_training_pipeline.py" \
+    "${parallel_args[@]}" \
+    "${model_args[@]}" \
+    "${dataset_args[@]}" \
+    "${training_args[@]}" \
+    "${optimizer_args[@]}" \
+    "${validation_args[@]}" \
+    "${rl_args[@]}" \
+    "${miscellaneous_args[@]}"

fastvideo-kernel/CMakeLists.txt

@@ -10,6 +10,8 @@ if(GPU_BACKEND STREQUAL "ROCM")
     enable_language(HIP)
 else()
     enable_language(CUDA)
+    # Ensure CUDA toolkit targets (CUDA::cudart, CUDA::cuda_driver, etc.) are available.
+    find_package(CUDAToolkit REQUIRED)
 endif()
 
 # Import common utils if needed, but we keep it simple for now
@@ -153,6 +155,30 @@ if(BUILD_CXX_KERNELS)
         $<$<COMPILE_LANGUAGE:CUDA>:${CUDA_FLAGS}>
     )
 
+    # Link against Torch libraries to avoid undefined symbols at import time
+    # (e.g., torch::autograd vtables) when loading the extension module.
+    target_link_libraries(fastvideo_kernel_ops PRIVATE ${TORCH_LIBRARIES})
+
+    # Also link against libtorch_python to satisfy Python-binding symbols
+    # (e.g., torch::PyWarningHandler) required by torch/extension.h.
+    execute_process(
+        COMMAND "${Python_EXECUTABLE}" -c "import torch; from pathlib import Path; p=Path(torch.__file__).parent/'lib'; m=sorted(p.glob('libtorch_python*')); print(str(m[0]) if m else '')"
+        OUTPUT_VARIABLE TORCH_PYTHON_LIBRARY_PATH
+        OUTPUT_STRIP_TRAILING_WHITESPACE
+        ERROR_QUIET
+    )
+    if(TORCH_PYTHON_LIBRARY_PATH)
+        message(STATUS "TORCH_PYTHON_LIBRARY_PATH: ${TORCH_PYTHON_LIBRARY_PATH}")
+        target_link_libraries(fastvideo_kernel_ops PRIVATE "${TORCH_PYTHON_LIBRARY_PATH}")
+    else()
+        message(WARNING "Could not locate libtorch_python; fastvideo_kernel_ops may fail to import.")
+    endif()
+
+    # Link CUDA runtime + driver explicitly (fixes missing symbols like cuGetErrorString at import time)
+    if(NOT GPU_BACKEND STREQUAL "ROCM")
+        target_link_libraries(fastvideo_kernel_ops PRIVATE CUDA::cudart CUDA::cuda_driver)
+    endif()
+
     # We install it to fastvideo_kernel/_C so we can load it to register the ops
     install(TARGETS fastvideo_kernel_ops LIBRARY DESTINATION fastvideo_kernel/_C)
 endif()

fastvideo-kernel/README.md

@@ -34,7 +34,7 @@ from fastvideo_kernel import sliding_tile_attention, video_sparse_attn, moba_att
 out = sliding_tile_attention(q, k, v, window_sizes, text_len)
 
 # Example: Video Sparse Attention (with Triton fallback)
-out = video_sparse_attn(q, k, v, block_sizes, topk=5)
+out = video_sparse_attn(q, k, v, block_sizes, block_sizes, topk=5)
 
 # Example: VMoBA
 out = moba_attn_varlen(q, k, v, cu_seqlens_q, cu_seqlens_k, ...)

fastvideo-kernel/csrc/attention/block_sparse_h100.cu

@@ -639,7 +639,8 @@ void bwd_attend_ker(const __grid_constant__ bwd_globals<D> g) {
 
     // store kq and vq
 
-    // ensuring all writes are finished
+    // ! the following two line seems unnecessary.
+    // tma::store_async_wait(); // ensure qg is finished
     __syncthreads();
 
     warpgroup::store(kg_smem[0], kg_reg);
@@ -660,145 +661,6 @@ void bwd_attend_ker(const __grid_constant__ bwd_globals<D> g) {
     tma::store_async_wait(); 
 }
 
-
-template<int D>
-void block_sparse_attention_forward_impl(
-    bf16* d_q, bf16* d_k, bf16* d_v, float* d_l, bf16* d_o,
-    int batch, int qo_heads, int kv_heads, int seq_len, int hr,
-    int max_kv_blocks_per_q, 
-    int32_t* q2k_block_sparse_index_ptr,
-    int32_t* q2k_block_sparse_num_ptr,
-    int32_t* block_size_ptr,
-    cudaStream_t stream
-) {
-    using K = fwd_attend_ker_tile_dims<D>;
-    using q_tile    =         st_bf<K::qo_height, K::tile_width>;
-    using k_tile    =         st_bf<K::kv_height, K::tile_width>;
-    using v_tile    =         st_bf<K::kv_height, K::tile_width>;
-    using l_col_vec = col_vec<st_fl<K::qo_height, K::tile_width>>;
-    using o_tile    =         st_bf<K::qo_height, K::tile_width>;
-
-    using q_global = gl<bf16,  -1, -1, -1, -1, q_tile>;
-    using k_global = gl<bf16,  -1, -1, -1, -1, k_tile>;
-    using v_global = gl<bf16,  -1, -1, -1, -1, v_tile>;
-    using l_global = gl<float, -1, -1, -1, -1, l_col_vec>;
-    using o_global = gl<bf16,  -1, -1, -1, -1, o_tile>;
-
-    using globals      = fwd_globals<D>;
-
-    q_global qg_arg{d_q, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    k_global kg_arg{d_k, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    v_global vg_arg{d_v, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    l_global lg_arg{d_l, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,   static_cast<unsigned int>(seq_len)};
-    o_global og_arg{d_o, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-
-    globals g{
-        qg_arg, kg_arg, vg_arg, lg_arg, og_arg, 
-        static_cast<int>(seq_len), static_cast<int>(hr), static_cast<int>(max_kv_blocks_per_q), 
-        q2k_block_sparse_index_ptr, q2k_block_sparse_num_ptr, block_size_ptr
-    };
-
-    // Shared memory size for the kernel
-    // 54000 bytes is calibrated for H100 shared memory constraints for these tile sizes
-    constexpr int mem_size = 54000;
-
-    dim3 grid(seq_len/(64), qo_heads, batch);
-
-    cudaFuncSetAttribute(
-        fwd_attend_ker<D>,
-        cudaFuncAttributeMaxDynamicSharedMemorySize,
-        mem_size
-    );
-
-    fwd_attend_ker<D><<<grid, (128), mem_size, stream>>>(g);
-}
-
-template<int D>
-void block_sparse_attention_backward_impl(
-    bf16* d_q, bf16* d_k, bf16* d_v, bf16* d_o, bf16* d_og, float* d_l, float* d_d, float* d_qg, float* d_kg, float* d_vg,
-    int batch, int qo_heads, int kv_heads, int seq_len, int hr, int max_q_blocks_per_kv,
-    int32_t* k2q_block_sparse_index_ptr,
-    int32_t* k2q_block_sparse_num_ptr,
-    int32_t* block_size_ptr,
-    cudaStream_t stream
-) {
-    using G = bwd_attend_ker_tile_dims<D>;
-    using og_tile = st_bf<4*16, D>;
-    using o_tile  = st_bf<4*16, D>;
-    using d_tile  = col_vec<st_fl<4*16, D>>;
-
-    using og_global = gl<bf16,  -1, -1, -1, -1, og_tile>;
-    using o_global  = gl<bf16,  -1, -1, -1, -1, o_tile>;
-    using d_global  = gl<float, -1, -1, -1, -1, d_tile>;
-
-    using prep_globals = bwd_prep_globals<D>;
-
-    constexpr int mem_size_prep = kittens::MAX_SHARED_MEMORY;
-    int threads_prep = PREP_NUM_WARPS * kittens::WARP_THREADS;
-    dim3 grid_bwd_prep(seq_len/(PREP_NUM_WARPS*kittens::TILE_ROW_DIM<bf16>*4), qo_heads, batch);
-
-    og_global prep_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    o_global  prep_o_arg {d_o,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    d_global  prep_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
-
-    prep_globals bwd_g{prep_og_arg, prep_o_arg, prep_d_arg};
-
-    cudaFuncSetAttribute(
-        bwd_attend_prep_ker<D>,
-        cudaFuncAttributeMaxDynamicSharedMemorySize,
-        mem_size_prep
-    );
-    bwd_attend_prep_ker<D><<<grid_bwd_prep, threads_prep, mem_size_prep, stream>>>(bwd_g);
-
-    using bwd_q_tile    =         st_bf<G::tile_h_qo, G::tile_width>;
-    using bwd_k_tile    =         st_bf<G::tile_h,    G::tile_width>;
-    using bwd_v_tile    =         st_bf<G::tile_h,    G::tile_width>;
-    using bwd_og_tile   =         st_bf<G::tile_h_qo, G::tile_width>;
-    using bwd_qg_tile   =         st_fl<G::tile_h_qo, G::tile_width>;
-    using bwd_kg_tile   =         st_fl<G::tile_h,    G::tile_width>;
-    using bwd_vg_tile   =         st_fl<G::tile_h,    G::tile_width>;
-    using bwd_l_tile    = row_vec<st_fl<G::tile_h_qo, G::tile_h>>;
-    using bwd_d_tile    = row_vec<st_fl<G::tile_h_qo, G::tile_h>>;
-
-    using bwd_q_global  = gl<bf16,  -1, -1, -1, -1, bwd_q_tile>;
-    using bwd_k_global  = gl<bf16,  -1, -1, -1, -1, bwd_k_tile>;
-    using bwd_v_global  = gl<bf16,  -1, -1, -1, -1, bwd_v_tile>;
-    using bwd_og_global = gl<bf16,  -1, -1, -1, -1, bwd_og_tile>;
-    using bwd_qg_global = gl<float, -1, -1, -1, -1, bwd_qg_tile>;
-    using bwd_kg_global = gl<float, -1, -1, -1, -1, bwd_kg_tile>;
-    using bwd_vg_global = gl<float, -1, -1, -1, -1, bwd_vg_tile>;
-    using bwd_l_global  = gl<float, -1, -1, -1, -1, bwd_l_tile>;
-    using bwd_d_global  = gl<float, -1, -1, -1, -1, bwd_d_tile>;
-
-    using bwd_global_args = bwd_globals<D>;
-
-    bwd_q_global  bwd_q_arg {d_q,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_k_global  bwd_k_arg {d_k,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_v_global  bwd_v_arg {d_v,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_og_global bwd_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_qg_global bwd_qg_arg{d_qg, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_kg_global bwd_kg_arg{d_kg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_vg_global bwd_vg_arg{d_vg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), static_cast<uint32_t>(D)};
-    bwd_l_global  bwd_l_arg {d_l,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
-    bwd_d_global  bwd_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
-
-    bwd_global_args bwd_global{bwd_q_arg, bwd_k_arg, bwd_v_arg, bwd_og_arg, bwd_qg_arg, bwd_kg_arg, bwd_vg_arg, bwd_l_arg, bwd_d_arg, 
-                        static_cast<int>(seq_len), static_cast<int>(hr), static_cast<int>(max_q_blocks_per_kv),
-                        k2q_block_sparse_index_ptr, k2q_block_sparse_num_ptr, block_size_ptr};
-
-    dim3 grid_bwd_main(seq_len/64, qo_heads, batch);
-    int threads_main = 128;
-    // Calibrated shared memory sizes for different head dimensions
-    int bwd_mem_size = (D == 64) ? 72000 : 113000;
-
-    cudaFuncSetAttribute(
-        bwd_attend_ker<D>,
-        cudaFuncAttributeMaxDynamicSharedMemorySize,
-        bwd_mem_size
-    );
-    bwd_attend_ker<D><<<grid_bwd_main, threads_main, bwd_mem_size, stream>>>(bwd_global);
-}
-
 #include "pyutils/torch_helpers.cuh"
 #include <ATen/cuda/CUDAContext.h>
 #include <iostream>
@@ -810,23 +672,32 @@ block_sparse_attention_forward(
     torch::Tensor v, 
     torch::Tensor q2k_block_sparse_index, 
     torch::Tensor q2k_block_sparse_num, 
-    torch::Tensor block_size
+    torch::Tensor kv_block_size
 )
 {
     CHECK_INPUT(q);
     CHECK_INPUT(k);
     CHECK_INPUT(v);
 
+    // q shape: (batch, qo_heads, q_seq_len, head_dim)
+    // k shape: (batch, kv_heads, kv_seq_len, head_dim)
+    // v shape: (batch, kv_heads, kv_seq_len, head_dim)
+    // q2k_block_sparse_index shape: (batch, qo_heads, num_q_blocks, max_kv_blocks_per_q)
+    // q2k_block_sparse_num shape: (batch, qo_heads, num_q_blocks)
+    // kv_block_size shape: (num_kv_blocks) This does not need other dimensions because across all batch/heads the padding is the same.
+    
     auto batch    = q.size(0);
-    auto seq_len  = q.size(2); 
+    auto q_seq_len  = q.size(2);
+    auto kv_seq_len = k.size(2);
     auto head_dim = q.size(3); 
     auto qo_heads = q.size(1);
     auto kv_heads = k.size(1);
     auto max_kv_blocks_per_q = q2k_block_sparse_index.size(3);
-    auto num_q_blocks = block_size.size(0);
+    auto num_q_blocks = q2k_block_sparse_index.size(2);
+    auto num_kv_blocks = kv_block_size.size(0);
     TORCH_CHECK(batch==1, "Batch size dim will be removed in the future, please set batch to 1");
-    TORCH_CHECK(num_q_blocks * 64 == seq_len, "This kernel supports variable block size, but it assumes the input sequence is properly padded.");
-    TORCH_CHECK(num_q_blocks == q2k_block_sparse_index.size(2), "Number of Q blocks does not match between q2k_block_sparse_index and block_size");
+    TORCH_CHECK(num_q_blocks * BLOCK_M == q_seq_len, "This kernel supports variable q block size, but it assumes the input sequence is properly padded.");
+    TORCH_CHECK(num_kv_blocks * BLOCK_M == kv_seq_len, "This kernel supports variable kv block size, but it assumes the input sequence is properly padded.");
     // check to see that these dimensions match for all inputs
     TORCH_CHECK(q.size(0) == batch, "Q batch dimension - idx 0 - must match for all inputs");
     TORCH_CHECK(k.size(0) == batch, "K batch dimension - idx 0 - must match for all inputs");
@@ -834,11 +705,9 @@ block_sparse_attention_forward(
     TORCH_CHECK(q2k_block_sparse_index.size(0) == batch, "q2k_block_sparse_index batch dimension - idx 0 - must match for all inputs");
     TORCH_CHECK(q2k_block_sparse_num.size(0) == batch, "q2k_block_sparse_num batch dimension - idx 0 - must match for all inputs");
 
-    TORCH_CHECK(q.size(2) == seq_len, "Q sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(k.size(2) == seq_len, "K sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(v.size(2) == seq_len, "V sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(q2k_block_sparse_index.size(2) == seq_len / BLOCK_M, "q2k_block_sparse_index idx 2 - must match seq_len / BLOCK_M");
-    TORCH_CHECK(q2k_block_sparse_num.size(2) == seq_len / BLOCK_M, "q2k_block_sparse_num idx 2 - must match seq_len / BLOCK_M");
+    TORCH_CHECK(v.size(2) == kv_seq_len, "V sequence length dimension - idx 2 - must match K inputs");
+    TORCH_CHECK(q2k_block_sparse_num.size(2) == num_q_blocks, "q2k_block_sparse_num idx 2 - must match num_q_blocks");
+
 
     TORCH_CHECK(q.size(3) == head_dim, "Q head dimension - idx 3 - must match for all non-vector inputs");
     TORCH_CHECK(k.size(3) == head_dim, "K head dimension - idx 3 - must match for all non-vector inputs");
@@ -864,12 +733,12 @@ block_sparse_attention_forward(
     // for the returned outputs
     torch::Tensor o     = torch::empty({static_cast<const uint>(batch), 
                                         static_cast<const uint>(qo_heads), 
-                                        static_cast<const uint>(seq_len), 
+                                        static_cast<const uint>(q_seq_len), 
                                         static_cast<const uint>(head_dim)}, v.options());
     
     torch::Tensor l_vec = torch::empty({static_cast<const uint>(batch), 
                                         static_cast<const uint>(qo_heads), 
-                                        static_cast<const uint>(seq_len), 
+                                        static_cast<const uint>(q_seq_len), 
                                         static_cast<const uint>(1)}, 
                                         torch::TensorOptions().dtype(torch::kFloat).device(q.device()).memory_format(at::MemoryFormat::Contiguous));
         
@@ -880,32 +749,110 @@ block_sparse_attention_forward(
     float* l_ptr = reinterpret_cast<float*>(l_vec.data_ptr<float>());
     float* d_l   = reinterpret_cast<float*>(l_ptr);
 
+    //cudadevicesynchronize();
     const c10::cuda::OptionalCUDAGuard device_guard(q.device());    
     const cudaStream_t stream = at::cuda::getCurrentCUDAStream().stream(); 
 
-    // Temporated implementation to avoid code duplication between head_dim=64 and 128
     if (head_dim == 64) {
-        block_sparse_attention_forward_impl<64>(
-            d_q, d_k, d_v, d_l, d_o,
-            batch, qo_heads, kv_heads, seq_len, hr,
-            max_kv_blocks_per_q,
-            reinterpret_cast<int32_t*>(q2k_block_sparse_index.data_ptr()),
+        using q_tile    =         st_bf<fwd_attend_ker_tile_dims<64>::qo_height, fwd_attend_ker_tile_dims<64>::tile_width>;
+        using k_tile    =         st_bf<fwd_attend_ker_tile_dims<64>::kv_height, fwd_attend_ker_tile_dims<64>::tile_width>;
+        using v_tile    =         st_bf<fwd_attend_ker_tile_dims<64>::kv_height, fwd_attend_ker_tile_dims<64>::tile_width>;
+        using l_col_vec = col_vec<st_fl<fwd_attend_ker_tile_dims<64>::qo_height, fwd_attend_ker_tile_dims<64>::tile_width>>;
+        using o_tile    =         st_bf<fwd_attend_ker_tile_dims<64>::qo_height, fwd_attend_ker_tile_dims<64>::tile_width>;
+
+        using q_global = gl<bf16,  -1, -1, -1, -1, q_tile>;
+        using k_global = gl<bf16,  -1, -1, -1, -1, k_tile>;
+        using v_global = gl<bf16,  -1, -1, -1, -1, v_tile>;
+        using l_global = gl<float, -1, -1, -1, -1, l_col_vec>;
+        using o_global = gl<bf16,  -1, -1, -1, -1, o_tile>;
+
+        using globals      = fwd_globals<64>;
+
+        q_global qg_arg{d_q, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+        k_global kg_arg{d_k, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 64U};
+        v_global vg_arg{d_v, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 64U};
+        l_global lg_arg{d_l, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,   static_cast<unsigned int>(q_seq_len)};
+        o_global og_arg{d_o, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+
+        globals g{
+            qg_arg, 
+            kg_arg, 
+            vg_arg, 
+            lg_arg, 
+            og_arg, 
+            static_cast<int>(q_seq_len), 
+            static_cast<int>(hr), 
+            static_cast<int>(max_kv_blocks_per_q), 
+            reinterpret_cast<int32_t*>(q2k_block_sparse_index.data_ptr()), 
             reinterpret_cast<int32_t*>(q2k_block_sparse_num.data_ptr()),
-            reinterpret_cast<int32_t*>(block_size.data_ptr()),
-            stream
+            reinterpret_cast<int32_t*>(kv_block_size.data_ptr())
+        };
+
+        constexpr int mem_size = 54000;
+
+        dim3 grid(q_seq_len/(BLOCK_M), qo_heads, batch);
+
+        cudaFuncSetAttribute(
+            fwd_attend_ker<64>,
+            cudaFuncAttributeMaxDynamicSharedMemorySize,
+            mem_size
         );
-    } else if (head_dim == 128) {
-        block_sparse_attention_forward_impl<128>(
-            d_q, d_k, d_v, d_l, d_o,
-            batch, qo_heads, kv_heads, seq_len, hr,
-            max_kv_blocks_per_q,
-            reinterpret_cast<int32_t*>(q2k_block_sparse_index.data_ptr()),
+
+        fwd_attend_ker<64><<<grid, (128), mem_size, stream>>>(g);
+
+        CHECK_CUDA_ERROR(cudaGetLastError());
+        // cudaStreamSynchronize(stream);
+    }
+
+    if (head_dim == 128) {
+        using q_tile    =         st_bf<fwd_attend_ker_tile_dims<128>::qo_height, fwd_attend_ker_tile_dims<128>::tile_width>;
+        using k_tile    =         st_bf<fwd_attend_ker_tile_dims<128>::kv_height, fwd_attend_ker_tile_dims<128>::tile_width>;
+        using v_tile    =         st_bf<fwd_attend_ker_tile_dims<128>::kv_height, fwd_attend_ker_tile_dims<128>::tile_width>;
+        using l_col_vec = col_vec<st_fl<fwd_attend_ker_tile_dims<128>::qo_height, fwd_attend_ker_tile_dims<128>::tile_width>>;
+        using o_tile    =         st_bf<fwd_attend_ker_tile_dims<128>::qo_height, fwd_attend_ker_tile_dims<128>::tile_width>;
+
+        using q_global = gl<bf16,  -1, -1, -1, -1, q_tile>;
+        using k_global = gl<bf16,  -1, -1, -1, -1, k_tile>;
+        using v_global = gl<bf16,  -1, -1, -1, -1, v_tile>;
+        using l_global = gl<float, -1, -1, -1, -1, l_col_vec>;
+        using o_global = gl<bf16,  -1, -1, -1, -1, o_tile>;
+
+        using globals      = fwd_globals<128>;
+
+        q_global qg_arg{d_q, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+        k_global kg_arg{d_k, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 128U};
+        v_global vg_arg{d_v, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 128U};
+        l_global lg_arg{d_l, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,   static_cast<unsigned int>(q_seq_len)};
+        o_global og_arg{d_o, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+
+        globals g{
+            qg_arg, 
+            kg_arg, 
+            vg_arg, 
+            lg_arg, 
+            og_arg, 
+            static_cast<int>(q_seq_len), 
+            static_cast<int>(hr), 
+            static_cast<int>(max_kv_blocks_per_q), 
+            reinterpret_cast<int32_t*>(q2k_block_sparse_index.data_ptr()), 
             reinterpret_cast<int32_t*>(q2k_block_sparse_num.data_ptr()),
-            reinterpret_cast<int32_t*>(block_size.data_ptr()),
-            stream
+            reinterpret_cast<int32_t*>(kv_block_size.data_ptr())
+        };
+
+        constexpr int mem_size = 54000;
+
+        dim3 grid(q_seq_len/(BLOCK_M), qo_heads, batch);
+
+        cudaFuncSetAttribute(
+            fwd_attend_ker<128>,
+            cudaFuncAttributeMaxDynamicSharedMemorySize,
+            mem_size
         );
-    } else {
-        TORCH_CHECK(false, "Unsupported head_dim: ", head_dim, ". Only 64 and 128 are supported.");
+
+        fwd_attend_ker<128><<<grid, (128), mem_size, stream>>>(g);
+
+        CHECK_CUDA_ERROR(cudaGetLastError());
+        // cudaStreamSynchronize(stream);
     }
 
     return {o, l_vec};
@@ -921,7 +868,7 @@ block_sparse_attention_backward(torch::Tensor q,
                    torch::Tensor og,
                    torch::Tensor k2q_block_sparse_index,
                    torch::Tensor k2q_block_sparse_num,
-                   torch::Tensor block_size)
+                   torch::Tensor kv_block_size)
 {
     CHECK_INPUT(q);
     CHECK_INPUT(k);
@@ -930,11 +877,23 @@ block_sparse_attention_backward(torch::Tensor q,
     CHECK_INPUT(o);
     CHECK_INPUT(og);
 
+    // q: [batch, qo_heads, q_seq_len, head_dim]
+    // k: [batch, kv_heads, kv_seq_len, head_dim]
+    // v: [batch, kv_heads, kv_seq_len, head_dim]
+    // o: [batch, qo_heads, q_seq_len, head_dim]
+    // l_vec: [batch, qo_heads, q_seq_len, 1]
+    // og: [batch, qo_heads, q_seq_len, head_dim]
+    // k2q_block_sparse_index: [batch, kv_heads, num_kv_blocks, max_num_q_blocks]
+    // k2q_block_sparse_num: [batch, kv_heads, num_kv_blocks]
+    // kv_block_size: [num_kv_blocks]
+
     auto batch    = q.size(0);
-    auto seq_len  = q.size(2);
+    auto q_seq_len  = q.size(2);
+    auto kv_seq_len = k.size(2);
     auto head_dim = q.size(3);
     auto max_q_blocks_per_kv = k2q_block_sparse_index.size(3);
-    TORCH_CHECK(k2q_block_sparse_index.size(2) == block_size.size(0), "k2q_block_sparse_index.size(2) must match block_size.size(0)");
+    auto num_kv_blocks = kv_block_size.size(0);
+    TORCH_CHECK(k2q_block_sparse_index.size(2) == num_kv_blocks, "k2q_block_sparse_index.size(2) must match num_kv_blocks (kv_block_size.size(0))");
     // check to see that these dimensions match for all inputs
     TORCH_CHECK(q.size(0)     == batch, "Q  batch dimension - idx 0 - must match for all inputs");
     TORCH_CHECK(k.size(0)     == batch, "K  batch dimension - idx 0 - must match for all inputs");
@@ -945,23 +904,18 @@ block_sparse_attention_backward(torch::Tensor q,
     TORCH_CHECK(k2q_block_sparse_index.size(0) == batch, "k2q_block_sparse_index batch dimension - idx 0 - must match for all inputs");
     TORCH_CHECK(k2q_block_sparse_num.size(0) == batch, "k2q_block_sparse_num batch dimension - idx 0 - must match for all inputs");
 
-
-    TORCH_CHECK(q.size(2)     == seq_len, "Q  sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(k.size(2)     == seq_len, "K  sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(v.size(2)     == seq_len, "V  sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(l_vec.size(2) == seq_len, "L  sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(o.size(2)     == seq_len, "O  sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(og.size(2)    == seq_len, "OG sequence length dimension - idx 2 - must match for all inputs");
-    TORCH_CHECK(k2q_block_sparse_index.size(2) == seq_len / BLOCK_N, "k2q_block_sparse_index idx 2 - must match seq_len / BLOCK_N");
-    TORCH_CHECK(k2q_block_sparse_num.size(2) == seq_len / BLOCK_N, "k2q_block_sparse_num idx 2 - must match seq_len / BLOCK_N");
+    TORCH_CHECK(v.size(2)     == kv_seq_len, "V  sequence length dimension - idx 2 - must match K sequence length");
+    TORCH_CHECK(l_vec.size(2) == q_seq_len, "L  sequence length dimension - idx 2 - must match Q sequence length");
+    TORCH_CHECK(o.size(2)     == q_seq_len, "O  sequence length dimension - idx 2 - must match Q sequence length");
+    TORCH_CHECK(og.size(2)    == q_seq_len, "OG sequence length dimension - idx 2 - must match Q sequence length");
+    TORCH_CHECK(k2q_block_sparse_index.size(2) == num_kv_blocks, "k2q_block_sparse_index idx 2 - must match num_kv_blocks (kv_block_size.size(0))");
+    TORCH_CHECK(k2q_block_sparse_num.size(2) == num_kv_blocks, "k2q_block_sparse_num idx 2 - must match num_kv_blocks (kv_block_size.size(0))");
 
     TORCH_CHECK(q.size(3)  == head_dim, "Q  head dimension - idx 3 - must match for all non-vector inputs");
     TORCH_CHECK(k.size(3)  == head_dim, "K  head dimension - idx 3 - must match for all non-vector inputs");
     TORCH_CHECK(v.size(3)  == head_dim, "V  head dimension - idx 3 - must match for all non-vector inputs");
     TORCH_CHECK(o.size(3)  == head_dim, "O  head dimension - idx 3 - must match for all non-vector inputs");
     TORCH_CHECK(og.size(3) == head_dim, "OG head dimension - idx 3 - must match for all non-vector inputs");
-    
-
 
     auto qo_heads = q.size(1);
     auto kv_heads = k.size(1);
@@ -988,20 +942,20 @@ block_sparse_attention_backward(torch::Tensor q,
 
     torch::Tensor qg = torch::zeros({static_cast<const uint>(batch), 
                                      static_cast<const uint>(qo_heads), 
-                                     static_cast<const uint>(seq_len), 
+                                     static_cast<const uint>(q_seq_len), 
                                      static_cast<const uint>(head_dim)},   l_vec.options());
     torch::Tensor kg = torch::zeros({static_cast<const uint>(batch), 
                                      static_cast<const uint>(kv_heads), 
-                                     static_cast<const uint>(seq_len), 
+                                     static_cast<const uint>(kv_seq_len), 
                                      static_cast<const uint>(head_dim)},   l_vec.options());
     torch::Tensor vg = torch::zeros({static_cast<const uint>(batch), 
                                      static_cast<const uint>(kv_heads), 
-                                     static_cast<const uint>(seq_len), 
+                                     static_cast<const uint>(kv_seq_len), 
                                      static_cast<const uint>(head_dim)},   l_vec.options());
     
     torch::Tensor d_vec = torch::empty({static_cast<const uint>(batch), 
                                         static_cast<const uint>(qo_heads), 
-                                        static_cast<const uint>(seq_len), 
+                                        static_cast<const uint>(q_seq_len), 
                                         static_cast<const uint>(1)},       l_vec.options());
 
     float*         qg_ptr = qg.data_ptr<float>();
@@ -1030,7 +984,7 @@ block_sparse_attention_backward(torch::Tensor q,
     //  cudaStreamSynchronize(stream);
 
     // TORCH_CHECK(seq_len % (4*kittens::TILE_DIM*4) == 0, "sequence length must be divisible by 256");
-    dim3 grid_bwd(seq_len/(PREP_NUM_WARPS*kittens::TILE_ROW_DIM<bf16>*4), qo_heads, batch);
+    dim3 grid_bwd(q_seq_len/(PREP_NUM_WARPS*kittens::TILE_ROW_DIM<bf16>*4), qo_heads, batch);
 
     if (head_dim == 64)  {
         using og_tile = st_bf<4*16, 64>;
@@ -1043,9 +997,9 @@ block_sparse_attention_backward(torch::Tensor q,
 
         using bwd_prep_globals = bwd_prep_globals<64>;
 
-        og_global prep_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 64U};
-        o_global  prep_o_arg {d_o,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 64U};
-        d_global  prep_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
+        og_global prep_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+        o_global  prep_o_arg {d_o,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+        d_global  prep_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(q_seq_len)};
 
         bwd_prep_globals bwd_g{prep_og_arg, prep_o_arg, prep_d_arg};
 
@@ -1082,15 +1036,15 @@ block_sparse_attention_backward(torch::Tensor q,
 
         using bwd_global_args = bwd_globals<64>;
 
-        bwd_q_global  bwd_q_arg {d_q,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_k_global  bwd_k_arg {d_k,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_v_global  bwd_v_arg {d_v,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_og_global bwd_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_qg_global bwd_qg_arg{d_qg, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_kg_global bwd_kg_arg{d_kg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_vg_global bwd_vg_arg{d_vg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 64U};
-        bwd_l_global  bwd_l_arg {d_l,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
-        bwd_d_global  bwd_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
+        bwd_q_global  bwd_q_arg {d_q,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+        bwd_k_global  bwd_k_arg {d_k,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 64U};
+        bwd_v_global  bwd_v_arg {d_v,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 64U};
+        bwd_og_global bwd_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+        bwd_qg_global bwd_qg_arg{d_qg, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 64U};
+        bwd_kg_global bwd_kg_arg{d_kg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 64U};
+        bwd_vg_global bwd_vg_arg{d_vg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 64U};
+        bwd_l_global  bwd_l_arg {d_l,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(q_seq_len)};
+        bwd_d_global  bwd_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(q_seq_len)};
 
         bwd_global_args bwd_global{bwd_q_arg, 
                         bwd_k_arg, 
@@ -1101,14 +1055,14 @@ block_sparse_attention_backward(torch::Tensor q,
                         bwd_vg_arg, 
                         bwd_l_arg, 
                         bwd_d_arg, 
-                        static_cast<int>(seq_len), 
+                        static_cast<int>(kv_seq_len), // N is not used in the kernel
                         static_cast<int>(hr),
                         static_cast<int>(max_q_blocks_per_kv),
                         reinterpret_cast<int32_t*>(k2q_block_sparse_index.data_ptr()),
                         reinterpret_cast<int32_t*>(k2q_block_sparse_num.data_ptr()),
-                        reinterpret_cast<int32_t*>(block_size.data_ptr())};
+                        reinterpret_cast<int32_t*>(kv_block_size.data_ptr())};
 
-        dim3 grid_bwd_2(seq_len/64, qo_heads, batch);
+        dim3 grid_bwd_2(kv_seq_len/BLOCK_N, qo_heads, batch);
         threads = 128;
 
         //cudadevicesynchronize();
@@ -1147,9 +1101,9 @@ block_sparse_attention_backward(torch::Tensor q,
 
         using bwd_prep_globals = bwd_prep_globals<128>;
 
-        og_global prep_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 128U};
-        o_global  prep_o_arg {d_o,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 128U};
-        d_global  prep_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
+        og_global prep_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+        o_global  prep_o_arg {d_o,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+        d_global  prep_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(q_seq_len)};
 
         bwd_prep_globals bwd_g{prep_og_arg, prep_o_arg, prep_d_arg};
 
@@ -1186,15 +1140,15 @@ block_sparse_attention_backward(torch::Tensor q,
 
         using bwd_global_args = bwd_globals<128>;
 
-        bwd_q_global  bwd_q_arg {d_q,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_k_global  bwd_k_arg {d_k,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_v_global  bwd_v_arg {d_v,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_og_global bwd_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_qg_global bwd_qg_arg{d_qg, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_kg_global bwd_kg_arg{d_kg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_vg_global bwd_vg_arg{d_vg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(seq_len), 128U};
-        bwd_l_global  bwd_l_arg {d_l,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
-        bwd_d_global  bwd_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(seq_len)};
+        bwd_q_global  bwd_q_arg {d_q,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+        bwd_k_global  bwd_k_arg {d_k,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 128U};
+        bwd_v_global  bwd_v_arg {d_v,  static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 128U};
+        bwd_og_global bwd_og_arg{d_og, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+        bwd_qg_global bwd_qg_arg{d_qg, static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), static_cast<unsigned int>(q_seq_len), 128U};
+        bwd_kg_global bwd_kg_arg{d_kg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 128U};
+        bwd_vg_global bwd_vg_arg{d_vg, static_cast<unsigned int>(batch), static_cast<unsigned int>(kv_heads), static_cast<unsigned int>(kv_seq_len), 128U};
+        bwd_l_global  bwd_l_arg {d_l,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(q_seq_len)};
+        bwd_d_global  bwd_d_arg {d_d,  static_cast<unsigned int>(batch), static_cast<unsigned int>(qo_heads), 1U,  static_cast<unsigned int>(q_seq_len)};
 
         bwd_global_args bwd_global{bwd_q_arg, 
                         bwd_k_arg, 
@@ -1205,14 +1159,14 @@ block_sparse_attention_backward(torch::Tensor q,
                         bwd_vg_arg, 
                         bwd_l_arg, 
                         bwd_d_arg, 
-                        static_cast<int>(seq_len), 
+                        static_cast<int>(kv_seq_len), // N is not used in the kernel
                         static_cast<int>(hr),
                         static_cast<int>(max_q_blocks_per_kv),
                         reinterpret_cast<int32_t*>(k2q_block_sparse_index.data_ptr()),
                         reinterpret_cast<int32_t*>(k2q_block_sparse_num.data_ptr()),
-                        reinterpret_cast<int32_t*>(block_size.data_ptr())};
+                        reinterpret_cast<int32_t*>(kv_block_size.data_ptr())};
 
-        dim3 grid_bwd_2(seq_len/64, qo_heads, batch);
+        dim3 grid_bwd_2(kv_seq_len/BLOCK_N, qo_heads, batch);
         threads = 128;
 
         //cudadevicesynchronize();
@@ -1233,4 +1187,4 @@ block_sparse_attention_backward(torch::Tensor q,
 
     return {qg, kg, vg};
     //cudadevicesynchronize();
-}
+}

fastvideo-kernel/csrc/turbodiffusion/norm/layernorm.cu

@@ -4,6 +4,7 @@
 #include <torch/all.h>
 #include <torch/python.h>
 #include <cutlass/cutlass.h>
+#include <cutlass/numeric_types.h>
 #include "common/common.hpp"
 #include "norm/layernorm.hpp"
 
@@ -14,10 +15,6 @@ auto layer_norm(
   std::optional<at::Tensor const> const B,
   std::optional<at::Tensor> Output
 ) {
-  using ElementIn = float;
-  using ElementOut = float;
-  using ElementWeight = float;
-
   int64_t const m = Input.size(0);
   int64_t const n = Input.size(1);
   torch::Device const input_device = Input.device();
@@ -26,31 +23,70 @@ auto layer_norm(
     Output.emplace(
       torch::empty(
         {m, n},
-        torch::TensorOptions().device(input_device).dtype(torch::kFloat32)
+        torch::TensorOptions().device(input_device).dtype(Input.scalar_type())
       )
     );
   }
 
+  TORCH_CHECK(Output.value().scalar_type() == Input.scalar_type(),
+              "Output dtype must match Input dtype. Got Output=",
+              Output.value().scalar_type(), ", Input=", Input.scalar_type());
+  if (W.has_value()) {
+    TORCH_CHECK(W.value().scalar_type() == Input.scalar_type(),
+                "W dtype must match Input dtype. Got W=",
+                W.value().scalar_type(), ", Input=", Input.scalar_type());
+  }
+  if (B.has_value()) {
+    TORCH_CHECK(B.value().scalar_type() == Input.scalar_type(),
+                "B dtype must match Input dtype. Got B=",
+                B.value().scalar_type(), ", Input=", Input.scalar_type());
+  }
 
   void *Iptr = Input.data_ptr();
   void *Wptr = W.has_value() ? W.value().data_ptr() : nullptr;
   void *Bptr = B.has_value() ? B.value().data_ptr() : nullptr;
   void *Optr = Output.value().data_ptr();
 
-  BOOL_SWITCH(B.has_value(), BIAS, [&]{
-    BOOL_SWITCH(W.has_value(), AFFINE, [&]{
-    CONFIG_SWITCH(n, [&]{
-    layernorm<
-    ElementIn, ElementOut, ElementWeight,
-    AFFINE, BIAS,
-    MAX_HIDDEN_SIZE, NUM_THR_PER_CTA> (
-      Iptr, Wptr, Bptr,
-      Optr, eps, m, n,
-      at::cuda::getCurrentCUDAStream().stream()
-        );
+  auto stream = at::cuda::getCurrentCUDAStream().stream();
+  if (Input.scalar_type() == at::kHalf) {
+    using ElementIn = cutlass::half_t;
+    using ElementOut = cutlass::half_t;
+    using ElementWeight = cutlass::half_t;
+    BOOL_SWITCH(B.has_value(), BIAS, [&]{
+      BOOL_SWITCH(W.has_value(), AFFINE, [&]{
+        CONFIG_SWITCH(n, [&]{
+          layernorm<ElementIn, ElementOut, ElementWeight, AFFINE, BIAS, MAX_HIDDEN_SIZE, NUM_THR_PER_CTA>(
+              Iptr, Wptr, Bptr, Optr, eps, m, n, stream);
+        });
       });
     });
-  });
+  } else if (Input.scalar_type() == at::kBFloat16) {
+    using ElementIn = cutlass::bfloat16_t;
+    using ElementOut = cutlass::bfloat16_t;
+    using ElementWeight = cutlass::bfloat16_t;
+    BOOL_SWITCH(B.has_value(), BIAS, [&]{
+      BOOL_SWITCH(W.has_value(), AFFINE, [&]{
+        CONFIG_SWITCH(n, [&]{
+          layernorm<ElementIn, ElementOut, ElementWeight, AFFINE, BIAS, MAX_HIDDEN_SIZE, NUM_THR_PER_CTA>(
+              Iptr, Wptr, Bptr, Optr, eps, m, n, stream);
+        });
+      });
+    });
+  } else if (Input.scalar_type() == at::kFloat) {
+    using ElementIn = float;
+    using ElementOut = float;
+    using ElementWeight = float;
+    BOOL_SWITCH(B.has_value(), BIAS, [&]{
+      BOOL_SWITCH(W.has_value(), AFFINE, [&]{
+        CONFIG_SWITCH(n, [&]{
+          layernorm<ElementIn, ElementOut, ElementWeight, AFFINE, BIAS, MAX_HIDDEN_SIZE, NUM_THR_PER_CTA>(
+              Iptr, Wptr, Bptr, Optr, eps, m, n, stream);
+        });
+      });
+    });
+  } else {
+    TORCH_CHECK(false, "Unsupported dtype for layer_norm_cuda: ", Input.scalar_type());
+  }
   
     
 

fastvideo-kernel/csrc/turbodiffusion/norm/layernorm.hpp

@@ -68,9 +68,22 @@ public:
     // mean reduction
     float u = _reduce_sum(x, shared_data) / params.n;
 
+    // IMPORTANT:
+    // Loader pads out-of-range lanes with 0. That is OK for the sum, but after
+    // subtracting mean, those padded lanes become -u and would incorrectly
+    // contribute to the variance. Mask them back to 0 before variance reduction.
+    // We launch exactly NumThrPerCta threads for a 1xMaxHiddenSize tile,
+    // so each thread is responsible for a contiguous chunk in N.
+    int thr_n_offset = tidx * NumElementPerThread;
     CUTLASS_PRAGMA_UNROLL
-    for (int i = 0; i < NumElementPerThread; ++i)
-      x[i] -= u;
+    for (int i = 0; i < NumElementPerThread; ++i) {
+      int idx = thr_n_offset + i;
+      if (idx < params.n) {
+        x[i] -= u;
+      } else {
+        x[i] = 0.f;
+      }
+    }
 
     __syncthreads();
     // var reduction

fastvideo-kernel/csrc/turbodiffusion/norm/rmsnorm.cu

@@ -4,6 +4,7 @@
 #include <torch/all.h>
 #include <torch/python.h>
 #include <cutlass/cutlass.h>
+#include <cutlass/numeric_types.h>
 #include <pybind11/pybind11.h>
 
 #include "common/common.hpp"
@@ -16,10 +17,6 @@ auto rms_norm(
   std::optional<at::Tensor>& Output
 ) {
   
-  using ElementIn = float;
-  using ElementOut = float;
-  using ElementWeight = float;
-  
   int64_t const m = Input.size(0);
   int64_t const n = Input.size(1);
   torch::Device const input_device = Input.device();
@@ -28,27 +25,51 @@ auto rms_norm(
     Output.emplace(
       torch::empty(
         {m, n},
-        torch::TensorOptions().device(input_device).dtype(torch::kFloat32)
+        torch::TensorOptions().device(input_device).dtype(Input.scalar_type())
       )
     );
   }
 
+  TORCH_CHECK(Output.value().scalar_type() == Input.scalar_type(),
+              "Output dtype must match Input dtype. Got Output=",
+              Output.value().scalar_type(), ", Input=", Input.scalar_type());
+  if (Weight.has_value()) {
+    TORCH_CHECK(Weight.value().scalar_type() == Input.scalar_type(),
+                "Weight dtype must match Input dtype. Got Weight=",
+                Weight.value().scalar_type(), ", Input=", Input.scalar_type());
+  }
+
   void *Iptr = Input.data_ptr();
   void *Wptr = Weight.has_value() ? Weight.value().data_ptr() : nullptr;
   void *Optr = Output.value().data_ptr();
 
-
-  CONFIG_SWITCH(n, [&]{
-    rmsnorm<
-    ElementIn, ElementOut, ElementWeight,
-    MAX_HIDDEN_SIZE, NUM_THR_PER_CTA
-    > (
-      Iptr, Wptr,
-      Optr, 
-      eps, m, n, 
-      at::cuda::getCurrentCUDAStream().stream()
-    );
-});
+  if (Input.scalar_type() == at::kHalf) {
+    using ElementIn = cutlass::half_t;
+    using ElementOut = cutlass::half_t;
+    using ElementWeight = cutlass::half_t;
+    CONFIG_SWITCH(n, [&]{
+      rmsnorm<ElementIn, ElementOut, ElementWeight, MAX_HIDDEN_SIZE, NUM_THR_PER_CTA>(
+          Iptr, Wptr, Optr, eps, m, n, at::cuda::getCurrentCUDAStream().stream());
+    });
+  } else if (Input.scalar_type() == at::kBFloat16) {
+    using ElementIn = cutlass::bfloat16_t;
+    using ElementOut = cutlass::bfloat16_t;
+    using ElementWeight = cutlass::bfloat16_t;
+    CONFIG_SWITCH(n, [&]{
+      rmsnorm<ElementIn, ElementOut, ElementWeight, MAX_HIDDEN_SIZE, NUM_THR_PER_CTA>(
+          Iptr, Wptr, Optr, eps, m, n, at::cuda::getCurrentCUDAStream().stream());
+    });
+  } else if (Input.scalar_type() == at::kFloat) {
+    using ElementIn = float;
+    using ElementOut = float;
+    using ElementWeight = float;
+    CONFIG_SWITCH(n, [&]{
+      rmsnorm<ElementIn, ElementOut, ElementWeight, MAX_HIDDEN_SIZE, NUM_THR_PER_CTA>(
+          Iptr, Wptr, Optr, eps, m, n, at::cuda::getCurrentCUDAStream().stream());
+    });
+  } else {
+    TORCH_CHECK(false, "Unsupported dtype for rms_norm_cuda: ", Input.scalar_type());
+  }
   
 
   return Output;

fastvideo-kernel/pyproject.toml

@@ -9,7 +9,7 @@ build-backend = "scikit_build_core.build"
 
 [project]
 name = "fastvideo-kernel"
-version = "0.2.1"
+version = "0.2.4"
 description = "Unified CUDA kernels for FastVideo"
 readme = "README.md"
 requires-python = ">=3.10"

fastvideo-kernel/python/fastvideo_kernel/block_sparse_attn.py

@@ -0,0 +1,298 @@
+from __future__ import annotations
+
+import os
+from typing import Tuple
+
+import torch
+
+
+def _get_sm90_ops():
+    try:
+        from fastvideo_kernel._C import fastvideo_kernel_ops  # type: ignore
+    except Exception:
+        return None, None
+    return (
+        getattr(fastvideo_kernel_ops, "block_sparse_fwd", None),
+        getattr(fastvideo_kernel_ops, "block_sparse_bwd", None),
+    )
+
+
+def _is_sm90() -> bool:
+    if not torch.cuda.is_available():
+        return False
+    major, minor = torch.cuda.get_device_capability(0)
+    return major == 9 and minor == 0
+
+
+def _force_triton() -> bool:
+    # Force Triton even on SM90 and even if the compiled extension is available.
+    # Useful for CI / debugging / parity testing.
+    return os.environ.get("FASTVIDEO_KERNEL_VSA_FORCE_TRITON", "0") == "1"
+
+
+def _map_to_index_torch(block_map: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
+    """
+    Pure-torch (no triton) conversion:
+      block_map: [B, H, Q, KV] bool  (or [H, Q, KV] which will be treated as B=1)
+      returns:
+        index: [B, H, Q, KV] int32 (packed KV indices, -1 padding)
+        num:   [B, H, Q] int32 (#kv blocks per q block)
+    """
+    if block_map.dim() == 3:
+        block_map = block_map.unsqueeze(0)
+    if block_map.dim() != 4:
+        raise ValueError(f"block_map must be [B,H,Q,KV] (or [H,Q,KV]), got shape={tuple(block_map.shape)}")
+    if block_map.dtype != torch.bool:
+        block_map = block_map.to(torch.bool)
+
+    B, H, Q, KV = block_map.shape
+    index = torch.full((B, H, Q, KV), -1, dtype=torch.int32, device=block_map.device)
+    num = torch.zeros((B, H, Q), dtype=torch.int32, device=block_map.device)
+
+    # Small sizes in practice (B=1, H<=16, Q/KV<=64), so a Python loop is fine.
+    for b in range(B):
+        for h in range(H):
+            for q in range(Q):
+                kv_idx = torch.nonzero(block_map[b, h, q], as_tuple=False).flatten().to(torch.int32)
+                n = int(kv_idx.numel())
+                if n:
+                    index[b, h, q, :n] = kv_idx
+                num[b, h, q] = n
+    return index, num
+
+
+@torch.library.custom_op(
+    "fastvideo_kernel::block_sparse_attn_triton",
+    mutates_args=(),
+    device_types="cuda",
+)
+def block_sparse_attn_triton(
+    q: torch.Tensor,
+    k: torch.Tensor,
+    v: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    q = q.contiguous()
+    k = k.contiguous()
+    v = v.contiguous()
+    block_map = block_map.to(torch.bool)
+    q2k_idx, q2k_num = _map_to_index_torch(block_map)
+
+    from fastvideo_kernel.triton_kernels.block_sparse_attn_triton import (  # local import
+        triton_block_sparse_attn_forward,
+    )
+
+    o, M = triton_block_sparse_attn_forward(q, k, v, q2k_idx, q2k_num, variable_block_sizes)
+    return o, M
+
+
+@torch.library.register_fake("fastvideo_kernel::block_sparse_attn_triton")
+def _block_sparse_attn_triton_fake(
+    q: torch.Tensor,
+    k: torch.Tensor,
+    v: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    o = torch.empty_like(q)
+    M = torch.empty((q.shape[0], q.shape[1], q.shape[2]), device=q.device, dtype=torch.float32)
+    return o, M
+
+
+@torch.library.custom_op(
+    "fastvideo_kernel::block_sparse_attn_backward_triton",
+    mutates_args=(),
+    device_types="cuda",
+)
+def block_sparse_attn_backward_triton(
+    grad_output: torch.Tensor,
+    q: torch.Tensor,
+    k: torch.Tensor,
+    v: torch.Tensor,
+    o: torch.Tensor,
+    M: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+    grad_output = grad_output.contiguous()
+    block_map = block_map.to(torch.bool)
+    q2k_idx, q2k_num = _map_to_index_torch(block_map)
+    k2q_idx, k2q_num = _map_to_index_torch(block_map.transpose(-1, -2).contiguous())
+
+    from fastvideo_kernel.triton_kernels.block_sparse_attn_triton import (  # local import
+        triton_block_sparse_attn_backward,
+    )
+
+    dq, dk, dv = triton_block_sparse_attn_backward(
+        grad_output, q, k, v, o, M, q2k_idx, q2k_num, k2q_idx, k2q_num, variable_block_sizes
+    )
+    return dq, dk, dv
+
+
+@torch.library.register_fake("fastvideo_kernel::block_sparse_attn_backward_triton")
+def _block_sparse_attn_backward_triton_fake(
+    grad_output: torch.Tensor,
+    q: torch.Tensor,
+    k: torch.Tensor,
+    v: torch.Tensor,
+    o: torch.Tensor,
+    M: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+    dq = torch.empty_like(q)
+    dk = torch.empty_like(k)
+    dv = torch.empty_like(v)
+    return dq, dk, dv
+
+
+def _backward_triton(ctx, grad_o, grad_M):
+    q, k, v, o, M, block_map, variable_block_sizes = ctx.saved_tensors
+    dq, dk, dv = block_sparse_attn_backward_triton(grad_o, q, k, v, o, M, block_map, variable_block_sizes)
+    return dq, dk, dv, None, None
+
+
+def _setup_context_triton(ctx, inputs, output):
+    q, k, v, block_map, variable_block_sizes = inputs
+    o, M = output
+    ctx.save_for_backward(q, k, v, o, M, block_map, variable_block_sizes)
+
+
+block_sparse_attn_triton.register_autograd(_backward_triton, setup_context=_setup_context_triton)
+
+
+@torch.library.custom_op(
+    "fastvideo_kernel::block_sparse_attn_sm90",
+    mutates_args=(),
+    device_types="cuda",
+)
+def block_sparse_attn_sm90(
+    q_padded: torch.Tensor,
+    k_padded: torch.Tensor,
+    v_padded: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    block_sparse_fwd, _ = _get_sm90_ops()
+    if block_sparse_fwd is None:
+        raise ImportError("fastvideo_kernel_ops.block_sparse_fwd is not available")
+
+    q_padded = q_padded.contiguous()
+    k_padded = k_padded.contiguous()
+    v_padded = v_padded.contiguous()
+    block_map = block_map.to(torch.bool)
+    q2k_idx, q2k_num = _map_to_index_torch(block_map)
+
+    o_padded, lse_padded = block_sparse_fwd(
+        q_padded, k_padded, v_padded, q2k_idx, q2k_num, variable_block_sizes.int()
+    )
+    return o_padded, lse_padded
+
+
+@torch.library.register_fake("fastvideo_kernel::block_sparse_attn_sm90")
+def _block_sparse_attn_sm90_fake(
+    q_padded: torch.Tensor,
+    k_padded: torch.Tensor,
+    v_padded: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    o = torch.empty_like(q_padded)
+    lse = torch.empty((q_padded.shape[0], q_padded.shape[1], q_padded.shape[2], 1), device=q_padded.device, dtype=torch.float32)
+    return o, lse
+
+
+@torch.library.custom_op(
+    "fastvideo_kernel::block_sparse_attn_backward_sm90",
+    mutates_args=(),
+    device_types="cuda",
+)
+def block_sparse_attn_backward_sm90(
+    grad_output_padded: torch.Tensor,
+    q_padded: torch.Tensor,
+    k_padded: torch.Tensor,
+    v_padded: torch.Tensor,
+    o_padded: torch.Tensor,
+    lse_padded: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+    _, block_sparse_bwd = _get_sm90_ops()
+    if block_sparse_bwd is None:
+        raise ImportError("fastvideo_kernel_ops.block_sparse_bwd is not available")
+
+    grad_output_padded = grad_output_padded.contiguous()
+    block_map = block_map.to(torch.bool)
+    k2q_idx, k2q_num = _map_to_index_torch(block_map.transpose(-1, -2).contiguous())
+
+    dq, dk, dv = block_sparse_bwd(
+        q_padded,
+        k_padded,
+        v_padded,
+        o_padded,
+        lse_padded,
+        grad_output_padded,
+        k2q_idx,
+        k2q_num,
+        variable_block_sizes.int(),
+    )
+    # C++ kernel returns fp32 grads; cast back to match PyTorch convention if needed
+    return dq.to(grad_output_padded.dtype), dk.to(grad_output_padded.dtype), dv.to(grad_output_padded.dtype)
+
+
+@torch.library.register_fake("fastvideo_kernel::block_sparse_attn_backward_sm90")
+def _block_sparse_attn_backward_sm90_fake(
+    grad_output_padded: torch.Tensor,
+    q_padded: torch.Tensor,
+    k_padded: torch.Tensor,
+    v_padded: torch.Tensor,
+    o_padded: torch.Tensor,
+    lse_padded: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
+    dq = torch.empty_like(q_padded)
+    dk = torch.empty_like(k_padded)
+    dv = torch.empty_like(v_padded)
+    return dq, dk, dv
+
+
+def _backward_sm90(ctx, grad_o, grad_lse):
+    q, k, v, o, lse, block_map, variable_block_sizes = ctx.saved_tensors
+    dq, dk, dv = block_sparse_attn_backward_sm90(
+        grad_o, q, k, v, o, lse, block_map, variable_block_sizes
+    )
+    return dq, dk, dv, None, None
+
+
+def _setup_context_sm90(ctx, inputs, output):
+    q, k, v, block_map, variable_block_sizes = inputs
+    o, lse = output
+    ctx.save_for_backward(q, k, v, o, lse, block_map, variable_block_sizes)
+
+
+block_sparse_attn_sm90.register_autograd(_backward_sm90, setup_context=_setup_context_sm90)
+
+
+def block_sparse_attn(
+    q: torch.Tensor,
+    k: torch.Tensor,
+    v: torch.Tensor,
+    block_map: torch.Tensor,
+    variable_block_sizes: torch.Tensor,
+) -> Tuple[torch.Tensor, torch.Tensor]:
+    """
+    Unified block-sparse attention op with autograd support.
+    - On SM90 with compiled extension present: uses fastvideo_kernel_ops.block_sparse_fwd/bwd.
+    - Otherwise: uses Triton implementation (requires q/k/v to have same padded length today).
+    """
+    block_sparse_fwd, block_sparse_bwd = _get_sm90_ops()
+    if (not _force_triton()) and _is_sm90() and (block_sparse_fwd is not None) and (block_sparse_bwd is not None):
+        return block_sparse_attn_sm90(q, k, v, block_map, variable_block_sizes)
+    # Triton path: generally assumes q/k/v share the same padded length
+    if q.shape[2] != k.shape[2] or q.shape[2] != v.shape[2]:
+        raise RuntimeError("Triton fallback requires q/k/v to have the same padded length.")
+    return block_sparse_attn_triton(q, k, v, block_map, variable_block_sizes)
+
+

fastvideo-kernel/python/fastvideo_kernel/ops.py

@@ -1,5 +1,6 @@
 import math
 import torch
+from .block_sparse_attn import block_sparse_attn
 from .triton_kernels.block_sparse_attn_triton import triton_block_sparse_attn_forward
 from .triton_kernels.st_attn_triton import sliding_tile_attention_triton
 from .triton_kernels.index import map_to_index
@@ -45,14 +46,22 @@ def sliding_tile_attention(
     flag = shape_map[seq_shape]
 
     for head_idx, (t, h, w) in enumerate(window_size):
+        # Per-head slices are not contiguous in the batch dimension when batch>1
+        # (they keep the original head-stride). The TK kernel assumes contiguous
+        # [B, H, S, D] layout, so we materialize a contiguous [B,1,S,D] view.
+        q_h = q[:, head_idx:head_idx + 1].contiguous()
+        k_h = k[:, head_idx:head_idx + 1].contiguous()
+        v_h = v[:, head_idx:head_idx + 1].contiguous()
+        o_h = torch.empty_like(q_h)
         sta_fwd(
-            q[:, head_idx:head_idx + 1], k[:, head_idx:head_idx + 1],
-            v[:, head_idx:head_idx + 1], output[:, head_idx:head_idx + 1],
+            q_h, k_h,
+            v_h, o_h,
             t, h, w, text_length, False, has_text, flag
         )
+        output[:, head_idx:head_idx + 1] = o_h
 
     if has_text:
-        sta_fwd(q, k, v, output, 3, 3, 3, text_length, True, True, flag)
+        sta_fwd(q.contiguous(), k.contiguous(), v.contiguous(), output, 3, 3, 3, text_length, True, True, flag)
 
     return output[:, :, :seq_length]
 
@@ -62,6 +71,7 @@ def video_sparse_attn(
     k: torch.Tensor,
     v: torch.Tensor,
     variable_block_sizes: torch.Tensor,
+    q_variable_block_sizes: torch.Tensor,
     topk: int,
     block_size: int | tuple = 64,
     compress_attn_weight: torch.Tensor = None,
@@ -70,14 +80,42 @@ def video_sparse_attn(
         block_size = (block_size, block_size, block_size)
 
     block_elements = block_size[0] * block_size[1] * block_size[2]
-    batch, heads, seq_len, dim = q.shape
+    batch, heads, q_seq_len, dim = q.shape
+    kv_seq_len = k.shape[2]
+    if v.shape[2] != kv_seq_len:
+        raise ValueError(
+            f"Expected k and v to have the same sequence length, got "
+            f"k.shape[2]={kv_seq_len}, v.shape[2]={v.shape[2]}"
+        )
+    if k.shape[0] != batch or v.shape[0] != batch or k.shape[1] != heads or v.shape[1] != heads:
+        raise ValueError("Expected q/k/v to have the same batch and head dimensions.")
+
+    if q_seq_len % block_elements != 0 or kv_seq_len % block_elements != 0:
+        raise ValueError(
+            f"q_seq_len and kv_seq_len must be divisible by block_elements={block_elements}, "
+            f"got q_seq_len={q_seq_len}, kv_seq_len={kv_seq_len}"
+        )
+    q_num_blocks = q_seq_len // block_elements
+    kv_num_blocks = kv_seq_len // block_elements
+
+    if variable_block_sizes.numel() != kv_num_blocks:
+        raise ValueError(
+            f"variable_block_sizes must have length kv_num_blocks={kv_num_blocks}, "
+            f"got {variable_block_sizes.numel()}"
+        )
+
+    if q_variable_block_sizes.numel() != q_num_blocks:
+        raise ValueError(
+            f"q_variable_block_sizes must have length q_num_blocks={q_num_blocks}, "
+            f"got {q_variable_block_sizes.numel()}"
+        )
 
     # Compression branch
-    q_c = q.view(batch, heads, seq_len // block_elements, block_elements, dim)
-    k_c = k.view(batch, heads, seq_len // block_elements, block_elements, dim)
-    v_c = v.view(batch, heads, seq_len // block_elements, block_elements, dim)
+    q_c = q.view(batch, heads, q_num_blocks, block_elements, dim)
+    k_c = k.view(batch, heads, kv_num_blocks, block_elements, dim)
+    v_c = v.view(batch, heads, kv_num_blocks, block_elements, dim)
 
-    q_c = (q_c.float().sum(dim=3) / variable_block_sizes.view(1, 1, -1, 1)).to(
+    q_c = (q_c.float().sum(dim=3) / q_variable_block_sizes.view(1, 1, -1, 1)).to(
         q.dtype)
     k_c = (k_c.float().sum(dim=3) / variable_block_sizes.view(1, 1, -1, 1)).to(
         k.dtype)
@@ -88,9 +126,9 @@ def video_sparse_attn(
     attn = torch.softmax(scores, dim=-1)
     out_c = torch.matmul(attn, v_c)
 
-    out_c = out_c.view(batch, heads, seq_len // block_elements, 1, dim)
+    out_c = out_c.view(batch, heads, q_num_blocks, 1, dim)
     out_c = out_c.repeat(1, 1, 1, block_elements,
-                         1).view(batch, heads, seq_len, dim)
+                         1).view(batch, heads, q_seq_len, dim)
 
     # Sparse branch
     topk_idx = torch.topk(scores, topk, dim=-1).indices
@@ -100,12 +138,17 @@ def video_sparse_attn(
     idx, num = map_to_index(mask)
     
     if block_sparse_fwd is not None:
-        out_s = block_sparse_fwd(
-            q, k, v, idx, num, variable_block_sizes.int()
-        )[0] # block_sparse_fwd returns vector<Tensor>
+        # Use autograd-enabled wrapper so backward works (and still uses SM90 kernel when available)
+        out_s = block_sparse_attn(q, k, v, mask, variable_block_sizes)[0]
     else:
-        out_s, _ = triton_block_sparse_attn_forward(q, k, v, idx, num,
-                                                    variable_block_sizes)
+        if q_seq_len != kv_seq_len:
+            raise RuntimeError(
+                "q/k have different lengths, but the compiled CUDA kernel (block_sparse_fwd) "
+                "is not available. The Triton fallback currently requires q and k/v to have "
+                "the same padded length."
+            )
+        # Triton-only forward (kept for environments without the wrapper deps)
+        out_s, _ = triton_block_sparse_attn_forward(q, k, v, idx, num, variable_block_sizes)
 
     if compress_attn_weight is not None:
         return out_c * compress_attn_weight + out_s

fastvideo-kernel/python/fastvideo_kernel/version.py

@@ -1 +1 @@
-__version__ = "0.2.1"
+__version__ = "0.2.4"

fastvideo-kernel/tests/test_vsa.py

@@ -42,37 +42,13 @@ def block_sparse_kernel_test(Q, K, V, block_sparse_mask, variable_block_sizes, q
     q_padded = vsa_pad(Q, q_non_pad_index, q_num_blocks, BLOCK_M)
     k_padded = vsa_pad(K, kv_non_pad_index, kv_num_blocks, BLOCK_M)
     v_padded = vsa_pad(V, kv_non_pad_index, kv_num_blocks, BLOCK_M)
-    # Use raw kernel or triton
-    try:
-        from fastvideo_kernel._C import fastvideo_kernel_ops
-        raw_kernel = getattr(fastvideo_kernel_ops, "block_sparse_fwd", None)
-    except ImportError:
-        raw_kernel = None
+    # Use autograd-enabled wrapper (internally dispatches to SM90 kernel or Triton)
+    from fastvideo_kernel.block_sparse_attn import block_sparse_attn
+    output_padded, _aux = block_sparse_attn(
+        q_padded, k_padded, v_padded, block_sparse_mask, variable_block_sizes
+    )
 
-    from fastvideo_kernel.triton_kernels.index import map_to_index
-
-    
-    # Convert mask to indices
-    # block_sparse_mask is [H, M, N] bool
-    # We need to map it to index.
-    # block_sparse_mask needs to be expanded/reshaped? 
-    # generate_block_sparse_mask_for_function returns [H, NumBlocksQ, NumBlocksKV]
-    
-    # Ops.py logic:
-    # mask = torch.zeros_like(scores, dtype=torch.bool).scatter_(-1, topk_idx, True)
-    # idx, num = map_to_index(mask)
-    
-    idx, num = map_to_index(block_sparse_mask.unsqueeze(0)) # Add batch dim [1, H, M, N]
-    
-    if raw_kernel:
-        out_s = raw_kernel(q_padded, k_padded, v_padded, idx, num, variable_block_sizes.int())
-        output = out_s[0]
-    else:
-        # Fallback to triton testing if C++ not available
-        from fastvideo_kernel.triton_kernels.block_sparse_attn_triton import triton_block_sparse_attn_forward
-        output, _ = triton_block_sparse_attn_forward(q_padded, k_padded, v_padded, idx, num, variable_block_sizes)
-
-    output = output[:, :, q_non_pad_index, :]
+    output = output_padded[:, :, q_non_pad_index, :]
     output.backward(dO)
     return output, Q.grad, K.grad, V.grad
 
@@ -264,7 +240,6 @@ def generate_error_graphs_qkdiff(h, d, error_mode='all'):
 
     print("-" * 150)
 
-@pytest.mark.skip()
 def test_video_sparse_attention_backward():
     if not torch.cuda.is_available():
         return

fastvideo-kernel/tests/test_vsa_forward.py

@@ -3,6 +3,7 @@ import sys
 from typing import Tuple
 
 import torch
+import pytest
 
 from .utils import (
     generate_block_sparse_mask_for_function,
@@ -57,23 +58,14 @@ def block_sparse_forward_test(
     k_padded = ref.vsa_pad(K, kv_non_pad_index, kv_num_blocks, BLOCK_M)
     v_padded = ref.vsa_pad(V, kv_non_pad_index, kv_num_blocks, BLOCK_M)
 
-    # Use raw kernel or triton
+    # Use autograd-enabled wrapper (internally dispatches SM90 C++ vs Triton)
+    from fastvideo_kernel.block_sparse_attn import block_sparse_attn
     try:
-        from fastvideo_kernel._C import fastvideo_kernel_ops
-        raw_kernel = getattr(fastvideo_kernel_ops, "block_sparse_fwd", None)
-    except ImportError:
-        raw_kernel = None
-        
-    from fastvideo_kernel.triton_kernels.index import map_to_index
-    idx, num = map_to_index(block_sparse_mask)
-    
-    if raw_kernel:
-        out_padded = raw_kernel(q_padded, k_padded, v_padded, idx, num, variable_block_sizes.int())[0]
-    else:
-        from fastvideo_kernel.triton_kernels.block_sparse_attn_triton import triton_block_sparse_attn_forward
-        out_padded, _ = triton_block_sparse_attn_forward(
-            q_padded, k_padded, v_padded, idx, num, variable_block_sizes
+        out_padded, _aux = block_sparse_attn(
+            q_padded, k_padded, v_padded, block_sparse_mask, variable_block_sizes
         )
+    except RuntimeError as e:
+        pytest.skip(str(e))
 
     # Remove padding on the query side
     out = out_padded[:, :, q_non_pad_index, :]
@@ -156,11 +148,6 @@ def run_forward_qk_diff(
 ) -> Tuple[float, float]:
     """
     Forward-only correctness test for the case S_q != S_kv.
-
-    NOTE:
-      - The Triton backend supports different Q/KV logical lengths via padding.
-      - The SM90 (H100) CUDA backend currently assumes the same number of blocks
-        for Q and KV, so we skip this test there.
     """
     assert torch.cuda.is_available(), "VSA kernels require CUDA"
 

fastvideo/attention/backends/video_sparse_attn.py

@@ -276,8 +276,9 @@ class VideoSparseAttentionImpl(AttentionImpl):
             query,
             key,
             value,
-            variable_block_sizes=attn_metadata.variable_block_sizes,
-            topk=cur_topk,
+            attn_metadata.variable_block_sizes,
+            attn_metadata.variable_block_sizes,
+            cur_topk,
             block_size=VSA_TILE_SIZE,
             compress_attn_weight=gate_compress).transpose(1, 2)
 

fastvideo/configs/models/dits/longcat.py

@@ -26,7 +26,6 @@ class LongCatVideoArchConfig(DiTArchConfig):
         default_factory=lambda: [is_longcat_blocks])
 
     # Parameter name mapping for weight conversion
-    # Maps original LongCat third_party names -> native FastVideo names
     param_names_mapping: dict = field(
         default_factory=lambda: {
             # Embedders

fastvideo/configs/models/encoders/__init__.py

@@ -7,10 +7,11 @@ from fastvideo.configs.models.encoders.clip import (
 from fastvideo.configs.models.encoders.llama import LlamaConfig
 from fastvideo.configs.models.encoders.t5 import T5Config, T5LargeConfig
 from fastvideo.configs.models.encoders.qwen2_5 import Qwen2_5_VLConfig
+from fastvideo.configs.models.encoders.reason1 import Reason1ArchConfig, Reason1Config
 
 __all__ = [
     "EncoderConfig", "TextEncoderConfig", "ImageEncoderConfig",
     "BaseEncoderOutput", "CLIPTextConfig", "CLIPVisionConfig",
     "WAN2_1ControlCLIPVisionConfig", "LlamaConfig", "T5Config", "T5LargeConfig",
-    "Qwen2_5_VLConfig"
+    "Qwen2_5_VLConfig", "Reason1ArchConfig", "Reason1Config"
 ]

fastvideo/configs/models/encoders/reason1.py

@@ -0,0 +1,72 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Config for Reason1 (Qwen2.5-VL) text encoder."""
+
+from dataclasses import dataclass, field
+from typing import Any
+
+from fastvideo.configs.models.encoders.base import TextEncoderArchConfig, TextEncoderConfig
+
+
+@dataclass
+class Reason1ArchConfig(TextEncoderArchConfig):
+    """Architecture settings (defaults match Qwen2.5-VL-7B-Instruct)."""
+
+    architectures: list[str] = field(
+        default_factory=lambda: ["Qwen2_5_VLForConditionalGeneration"])
+    model_type: str = "qwen2_5_vl"
+
+    vocab_size: int = 152064
+    hidden_size: int = 3584
+    num_hidden_layers: int = 28
+    num_attention_heads: int = 28
+    num_key_value_heads: int = 4
+    intermediate_size: int = 18944
+
+    text_len: int = 512
+    hidden_state_skip_layer: int = 0
+    bos_token_id: int = 151643
+    pad_token_id: int = 151643
+    eos_token_id: int = 151645
+
+    image_token_id: int = 151655
+    video_token_id: int = 151656
+    vision_token_id: int = 151654
+    vision_start_token_id: int = 151652
+    vision_end_token_id: int = 151653
+
+    vision_config: dict[str, Any] | None = None
+
+    rope_theta: float = 1000000.0
+    rope_scaling: dict[str, Any] | None = field(default_factory=lambda: {
+        "type": "mrope",
+        "mrope_section": [16, 24, 24]
+    })
+    max_position_embeddings: int = 128000
+    max_window_layers: int = 28
+
+    embedding_concat_strategy: str = "mean_pooling"
+    n_layers_per_group: int = 5
+    num_embedding_padding_tokens: int = 512
+
+    attention_dropout: float = 0.0
+    hidden_act: str = "silu"
+    initializer_range: float = 0.02
+    rms_norm_eps: float = 1e-6
+
+    use_sliding_window: bool = False
+    sliding_window: int = 32768
+
+    tie_word_embeddings: bool = False
+    use_cache: bool = False
+    output_hidden_states: bool = True
+
+    torch_dtype: str = "bfloat16"
+    _attn_implementation: str = "flash_attention_2"
+
+
+@dataclass
+class Reason1Config(TextEncoderConfig):
+    """Reason1 text encoder config."""
+
+    arch_config: Reason1ArchConfig = field(default_factory=Reason1ArchConfig)
+    tokenizer_type: str = "Qwen/Qwen2.5-VL-7B-Instruct"

fastvideo/configs/models/vaes/__init__.py

@@ -1,4 +1,5 @@
 from fastvideo.configs.models.vaes.cosmosvae import CosmosVAEConfig
+from fastvideo.configs.models.vaes.cosmos2_5vae import Cosmos25VAEConfig
 from fastvideo.configs.models.vaes.hunyuanvae import HunyuanVAEConfig
 from fastvideo.configs.models.vaes.hunyuan15vae import Hunyuan15VAEConfig
 from fastvideo.configs.models.vaes.stepvideovae import StepVideoVAEConfig
@@ -9,5 +10,6 @@ __all__ = [
     "WanVAEConfig",
     "StepVideoVAEConfig",
     "CosmosVAEConfig",
+    "Cosmos25VAEConfig",
     "Hunyuan15VAEConfig",
 ]

fastvideo/configs/models/vaes/cosmos2_5vae.py

@@ -0,0 +1,223 @@
+"""Cosmos 2.5 (Wan2.1-style) VAE config and checkpoint-key mapping."""
+
+from __future__ import annotations
+
+import re
+from dataclasses import dataclass, field
+
+import torch
+
+from fastvideo.configs.models.vaes.base import VAEArchConfig, VAEConfig
+
+
+@dataclass
+class Cosmos25VAEArchConfig(VAEArchConfig):
+    _name_or_path: str = ""
+    base_dim: int = 96
+    decoder_base_dim: int | None = None
+    z_dim: int = 16
+    dim_mult: tuple[int, ...] = (1, 2, 4, 4)
+    num_res_blocks: int = 2
+    attn_scales: tuple[float, ...] = ()
+    temperal_downsample: tuple[bool, ...] = (False, True, True)
+    dropout: float = 0.0
+    is_residual: bool = False
+    in_channels: int = 3
+    out_channels: int = 3
+    patch_size: int | None = None
+    scale_factor_temporal: int = 4
+    scale_factor_spatial: int = 8
+    clip_output: bool = True
+
+    latents_mean: tuple[float, ...] = (
+        -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,
+    )
+    latents_std: tuple[float, ...] = (
+        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,
+    )
+
+    # Simple 1:1 renames. More complex decoder remapping is handled by
+    # `map_official_key()`.
+    param_names_mapping: dict[str, str] = field(
+        default_factory=lambda: {
+            r"^conv1\.(.*)$": r"quant_conv.\1",
+            r"^conv2\.(.*)$": r"post_quant_conv.\1",
+            r"^encoder\.conv1\.(.*)$": r"encoder.conv_in.\1",
+            r"^decoder\.conv1\.(.*)$": r"decoder.conv_in.\1",
+            r"^encoder\.head\.0\.gamma$": r"encoder.norm_out.gamma",
+            r"^encoder\.head\.2\.(.*)$": r"encoder.conv_out.\1",
+            r"^decoder\.head\.0\.gamma$": r"decoder.norm_out.gamma",
+            r"^decoder\.head\.2\.(.*)$": r"decoder.conv_out.\1",
+        })
+
+    @staticmethod
+    def map_official_key(key: str) -> str | None:
+        """Map a single official checkpoint key into FastVideo key space."""
+
+        def map_residual_subkey(prefix: str, sub: str) -> str | None:
+            if re.match(r"^residual\.0\.gamma$", sub):
+                return f"{prefix}.norm1.gamma"
+            m = re.match(r"^residual\.2\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.conv1.{m.group(1)}"
+            if re.match(r"^residual\.3\.gamma$", sub):
+                return f"{prefix}.norm2.gamma"
+            m = re.match(r"^residual\.6\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.conv2.{m.group(1)}"
+            m = re.match(r"^shortcut\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.conv_shortcut.{m.group(1)}"
+            return None
+
+        def map_attn_subkey(prefix: str, sub: str) -> str | None:
+            if re.match(r"^norm\.gamma$", sub):
+                return f"{prefix}.norm.gamma"
+            m = re.match(r"^to_qkv\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.to_qkv.{m.group(1)}"
+            m = re.match(r"^proj\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.proj.{m.group(1)}"
+            return None
+
+        def map_resample_subkey(prefix: str, sub: str) -> str | None:
+            m = re.match(r"^resample\.1\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.resample.1.{m.group(1)}"
+            m = re.match(r"^time_conv\.(weight|bias)$", sub)
+            if m:
+                return f"{prefix}.time_conv.{m.group(1)}"
+            return None
+
+        m = re.match(r"^conv1\.(weight|bias)$", key)
+        if m:
+            return f"quant_conv.{m.group(1)}"
+        m = re.match(r"^conv2\.(weight|bias)$", key)
+        if m:
+            return f"post_quant_conv.{m.group(1)}"
+        m = re.match(r"^(encoder|decoder)\.conv1\.(weight|bias)$", key)
+        if m:
+            return f"{m.group(1)}.conv_in.{m.group(2)}"
+        m = re.match(r"^(encoder|decoder)\.head\.0\.gamma$", key)
+        if m:
+            return f"{m.group(1)}.norm_out.gamma"
+        m = re.match(r"^(encoder|decoder)\.head\.2\.(weight|bias)$", key)
+        if m:
+            return f"{m.group(1)}.conv_out.{m.group(2)}"
+
+        m = re.match(r"^(encoder|decoder)\.middle\.0\.(.*)$", key)
+        if m:
+            return map_residual_subkey(f"{m.group(1)}.mid_block.resnets.0",
+                                       m.group(2))
+        m = re.match(r"^(encoder|decoder)\.middle\.1\.(.*)$", key)
+        if m:
+            return map_attn_subkey(f"{m.group(1)}.mid_block.attentions.0",
+                                   m.group(2))
+        m = re.match(r"^(encoder|decoder)\.middle\.2\.(.*)$", key)
+        if m:
+            return map_residual_subkey(f"{m.group(1)}.mid_block.resnets.1",
+                                       m.group(2))
+
+        m = re.match(r"^encoder\.downsamples\.(\d+)\.(.*)$", key)
+        if m:
+            idx = int(m.group(1))
+            sub = m.group(2)
+            if sub.startswith("residual.") or sub.startswith("shortcut."):
+                return map_residual_subkey(f"encoder.down_blocks.{idx}", sub)
+            if sub.startswith("resample.") or sub.startswith("time_conv."):
+                return map_resample_subkey(f"encoder.down_blocks.{idx}", sub)
+            return None
+
+        m = re.match(r"^decoder\.upsamples\.(\d+)\.(.*)$", key)
+        if m:
+            uidx = int(m.group(1))
+            sub = m.group(2)
+
+            if uidx in (0, 1, 2):
+                block_i, res_i = 0, uidx
+            elif uidx == 3:
+                block_i, res_i = 0, None
+            elif uidx in (4, 5, 6):
+                block_i, res_i = 1, uidx - 4
+            elif uidx == 7:
+                block_i, res_i = 1, None
+            elif uidx in (8, 9, 10):
+                block_i, res_i = 2, uidx - 8
+            elif uidx == 11:
+                block_i, res_i = 2, None
+            elif uidx in (12, 13, 14):
+                block_i, res_i = 3, uidx - 12
+            else:
+                return None
+
+            if res_i is None:
+                return map_resample_subkey(
+                    f"decoder.up_blocks.{block_i}.upsamplers.0",
+                    sub,
+                )
+
+            return map_residual_subkey(
+                f"decoder.up_blocks.{block_i}.resnets.{res_i}",
+                sub,
+            )
+
+        return None
+
+    temporal_compression_ratio: int = 4
+    spatial_compression_ratio: int = 8
+
+    def __post_init__(self):
+        self.scaling_factor: torch.Tensor = 1.0 / torch.tensor(
+            self.latents_std).view(1, self.z_dim, 1, 1, 1)
+        self.shift_factor: torch.Tensor = torch.tensor(self.latents_mean).view(
+            1, self.z_dim, 1, 1, 1)
+        self.temporal_compression_ratio = self.scale_factor_temporal
+        self.spatial_compression_ratio = self.scale_factor_spatial
+
+
+@dataclass
+class Cosmos25VAEConfig(VAEConfig):
+    """Cosmos2.5 VAE config."""
+
+    arch_config: Cosmos25VAEArchConfig = field(
+        default_factory=Cosmos25VAEArchConfig)
+
+    use_feature_cache: bool = True
+    use_tiling: bool = False
+    use_temporal_tiling: bool = False
+    use_parallel_tiling: bool = False
+
+    def __post_init__(self):
+        self.blend_num_frames = (self.tile_sample_min_num_frames -
+                                 self.tile_sample_stride_num_frames) * 2

fastvideo/configs/pipelines/__init__.py

@@ -1,6 +1,7 @@
 from fastvideo.configs.pipelines.base import (PipelineConfig,
                                               SlidingTileAttnConfig)
 from fastvideo.configs.pipelines.cosmos import CosmosConfig
+from fastvideo.configs.pipelines.cosmos2_5 import Cosmos25Config
 from fastvideo.configs.pipelines.hunyuan import FastHunyuanConfig, HunyuanConfig
 from fastvideo.configs.pipelines.hunyuan15 import Hunyuan15T2V480PConfig, Hunyuan15T2V720PConfig
 from fastvideo.configs.pipelines.registry import (
@@ -15,5 +16,5 @@ __all__ = [
     "Hunyuan15T2V480PConfig", "Hunyuan15T2V720PConfig", "SlidingTileAttnConfig",
     "WanT2V480PConfig", "WanI2V480PConfig", "WanT2V720PConfig",
     "WanI2V720PConfig", "StepVideoT2VConfig", "SelfForcingWanT2V480PConfig",
-    "CosmosConfig", "get_pipeline_config_cls_from_name"
+    "CosmosConfig", "Cosmos25Config", "get_pipeline_config_cls_from_name"
 ]

fastvideo/configs/pipelines/cosmos2_5.py

@@ -0,0 +1,89 @@
+# SPDX-License-Identifier: Apache-2.0
+from collections.abc import Callable
+from dataclasses import dataclass, field
+
+import torch
+
+from fastvideo.configs.models import DiTConfig, EncoderConfig, VAEConfig
+from fastvideo.configs.models.dits import Cosmos25VideoConfig
+from fastvideo.configs.models.dits.cosmos2_5 import Cosmos25ArchConfig
+from fastvideo.configs.models.encoders import BaseEncoderOutput
+from fastvideo.configs.models.encoders.reason1 import Reason1Config, Reason1ArchConfig
+from fastvideo.configs.models.vaes import Cosmos25VAEConfig
+from fastvideo.configs.pipelines.base import PipelineConfig, STA_Mode
+
+
+def _identity_preprocess_text(prompt: str) -> str:
+    return prompt
+
+
+def reason1_postprocess_text(outputs: BaseEncoderOutput) -> torch.Tensor:
+    hidden_states = getattr(outputs, "hidden_states", None)
+    if hidden_states is None:
+        raise ValueError("Reason1 postprocess requires outputs.hidden_states")
+
+    hs = list(hidden_states)[1:]
+    normed = []
+    for h in hs:
+        h = h.float()
+        h = (h - h.mean(dim=-1, keepdim=True)) / (h.std(dim=-1, keepdim=True) +
+                                                  1e-8)
+        normed.append(h)
+    return torch.cat(normed, dim=-1).to(hidden_states[0].dtype)
+
+
+@dataclass
+class Cosmos25Config(PipelineConfig):
+    """Configuration for Cosmos 2.5 (Predict2.5) video generation pipeline."""
+
+    dit_config: DiTConfig = field(default_factory=lambda: Cosmos25VideoConfig(
+        arch_config=Cosmos25ArchConfig(
+            num_attention_heads=16,
+            attention_head_dim=128,
+            in_channels=16,
+            out_channels=16,
+            num_layers=28,
+            patch_size=[1, 2, 2],
+            max_size=[128, 240, 240],
+            rope_scale=[1.0, 3.0, 3.0],
+            text_embed_dim=1024,
+            mlp_ratio=4.0,
+            adaln_lora_dim=256,
+            use_adaln_lora=True,
+            concat_padding_mask=True,
+            extra_pos_embed_type=None,
+            use_crossattn_projection=True,
+            rope_enable_fps_modulation=False,
+            qk_norm="rms_norm",
+        )))
+
+    vae_config: VAEConfig = field(default_factory=Cosmos25VAEConfig)
+
+    text_encoder_configs: tuple[EncoderConfig, ...] = field(
+        default_factory=lambda: (Reason1Config(arch_config=Reason1ArchConfig(
+            embedding_concat_strategy="full_concat")), ))
+
+    preprocess_text_funcs: tuple[Callable[[str], str], ...] = field(
+        default_factory=lambda: (_identity_preprocess_text, ))
+    postprocess_text_funcs: tuple[Callable[[BaseEncoderOutput], torch.Tensor],
+                                  ...] = field(default_factory=lambda:
+                                               (reason1_postprocess_text, ))
+
+    dit_precision: str = "bf16"
+    vae_precision: str = "bf16"
+    text_encoder_precisions: tuple[str, ...] = field(
+        default_factory=lambda: ("bf16", ))
+
+    embedded_cfg_scale: float = 0.0
+    flow_shift: float = 5.0
+
+    vae_tiling: bool = False
+    vae_sp: bool = False
+
+    STA_mode: STA_Mode = STA_Mode.NONE
+    skip_time_steps: int = 0
+
+    def __post_init__(self):
+        self.vae_config.load_encoder = True
+        self.vae_config.load_decoder = True
+        self._vae_latent_dim = 16

fastvideo/configs/pipelines/longcat.py

@@ -17,11 +17,7 @@ from fastvideo.configs.pipelines.base import PipelineConfig
 
 @dataclass
 class LongCatDiTArchConfig(DiTArchConfig):
-    """Extended DiTArchConfig with LongCat-specific fields.
-    
-    NOTE: This is for Phase 1 wrapper compatibility. For native model (Phase 2),
-    use LongCatVideoConfig from fastvideo.configs.models.dits.longcat instead.
-    """
+    """Extended DiTArchConfig with LongCat-specific fields."""
     # LongCat-specific architecture parameters
     adaln_tembed_dim: int = 512
     caption_channels: int = 4096
@@ -88,20 +84,16 @@ def umt5_postprocess_text(outputs: BaseEncoderOutput) -> torch.Tensor:
 
 @dataclass
 class LongCatT2V480PConfig(PipelineConfig):
-    """Configuration for LongCat pipeline (480p) aligned to LongCat-Video modules.
+    """Configuration for LongCat pipeline (480p).
 
     Components expected by loaders:
       - tokenizer: AutoTokenizer
       - text_encoder: UMT5EncoderModel
-      - transformer: LongCatVideoTransformer3DModel (Phase 1 wrapper)
-                  OR LongCatTransformer3DModel (Phase 2 native)
+      - transformer: LongCatTransformer3DModel
       - vae: AutoencoderKLWan (Wan VAE, 4x8 compression)
       - scheduler: FlowMatchEulerDiscreteScheduler
     """
 
-    # DiT config with LongCat-specific arch_config
-    # NOTE: For Phase 1 wrapper, uses LongCatDiTArchConfig
-    # For Phase 2 native model, can use LongCatVideoConfig directly
     dit_config: DiTConfig = field(
         default_factory=lambda: DiTConfig(arch_config=LongCatDiTArchConfig()))
 

fastvideo/configs/pipelines/registry.py

@@ -6,10 +6,14 @@ from collections.abc import Callable
 
 from fastvideo.configs.pipelines.base import PipelineConfig
 from fastvideo.configs.pipelines.cosmos import CosmosConfig
+from fastvideo.configs.pipelines.cosmos2_5 import Cosmos25Config
 from fastvideo.configs.pipelines.hunyuan import FastHunyuanConfig, HunyuanConfig
 from fastvideo.configs.pipelines.hunyuan15 import Hunyuan15T2V480PConfig, Hunyuan15T2V720PConfig
 from fastvideo.configs.pipelines.stepvideo import StepVideoT2VConfig
 from fastvideo.configs.pipelines.longcat import LongCatT2V480PConfig
+from fastvideo.configs.pipelines.turbodiffusion import (
+    TurboDiffusionT2V_1_3B_Config, TurboDiffusionT2V_14B_Config,
+    TurboDiffusionI2V_A14B_Config)
 
 # isort: off
 from fastvideo.configs.pipelines.wan import (
@@ -52,14 +56,29 @@ PIPE_NAME_TO_CONFIG: dict[str, type[PipelineConfig]] = {
     "Wan-AI/Wan2.2-T2V-A14B-Diffusers": Wan2_2_T2V_A14B_Config,
     "Wan-AI/Wan2.2-I2V-A14B-Diffusers": Wan2_2_I2V_A14B_Config,
     "nvidia/Cosmos-Predict2-2B-Video2World": CosmosConfig,
+    "KyleShao/Cosmos-Predict2.5-2B-Diffusers": Cosmos25Config,
     "FastVideo/Matrix-Game-2.0-Base-Diffusers": MatrixGameI2V480PConfig,
     "FastVideo/Matrix-Game-2.0-GTA-Diffusers": MatrixGameI2V480PConfig,
     "FastVideo/Matrix-Game-2.0-TempleRun-Diffusers": MatrixGameI2V480PConfig,
+    # LongCat Video models
+    "FastVideo/LongCat-Video-T2V-Diffusers": LongCatT2V480PConfig,
+    "FastVideo/LongCat-Video-I2V-Diffusers": LongCatT2V480PConfig,
+    "FastVideo/LongCat-Video-VC-Diffusers": LongCatT2V480PConfig,
+    # TurboDiffusion models
+    "loayrashid/TurboWan2.1-T2V-1.3B-Diffusers": TurboDiffusionT2V_1_3B_Config,
+    "loayrashid/TurboWan2.1-T2V-14B-Diffusers": TurboDiffusionT2V_14B_Config,
+    "loayrashid/TurboWan2.2-I2V-A14B-Diffusers": TurboDiffusionI2V_A14B_Config,
     # Add other specific weight variants
 }
 
 # For determining pipeline type from model ID
 PIPELINE_DETECTOR: dict[str, Callable[[str], bool]] = {
+    "longcatimagetovideo":
+    lambda id: "longcatimagetovideo" in id.lower(),
+    "longcatvideocontinuation":
+    lambda id: "longcatvideocontinuation" in id.lower(),
+    "longcat":
+    lambda id: "longcat" in id.lower(),
     "hunyuan":
     lambda id: "hunyuan" in id.lower(),
     "hunyuan15":
@@ -77,15 +96,21 @@ PIPELINE_DETECTOR: dict[str, Callable[[str], bool]] = {
     "stepvideo":
     lambda id: "stepvideo" in id.lower(),
     "cosmos":
-    lambda id: "cosmos" in id.lower(),
-    "longcat":
-    lambda id: "longcat" in id.lower(),
+    lambda id: "cosmos" in id.lower() and ("2.5" not in id.lower(
+    ) and "2_5" not in id.lower() and "25" not in id.lower()),
+    "cosmos25":
+    lambda id: "cosmos25" in id.lower(),
+    "turbodiffusion":
+    lambda id: "turbodiffusion" in id.lower() or "turbowan" in id.lower(),
     # Add other pipeline architecture detectors
 }
 
 # Fallback configs when exact match isn't found but architecture is detected
 PIPELINE_FALLBACK_CONFIG: dict[str, type[PipelineConfig]] = {
+    "longcatimagetovideo": LongCatT2V480PConfig,
+    "longcatvideocontinuation": LongCatT2V480PConfig,
     "longcat": LongCatT2V480PConfig,
+    "cosmos25": Cosmos25Config,
     "hunyuan":
     HunyuanConfig,  # Base Hunyuan config as fallback for any Hunyuan variant
     "matrixgame": MatrixGameI2V480PConfig,
@@ -96,7 +121,8 @@ PIPELINE_FALLBACK_CONFIG: dict[str, type[PipelineConfig]] = {
     "wanimagetovideo": WanI2V480PConfig,
     "wandmdpipeline": FastWan2_1_T2V_480P_Config,
     "wancausaldmdpipeline": SelfForcingWanT2V480PConfig,
-    "stepvideo": StepVideoT2VConfig
+    "stepvideo": StepVideoT2VConfig,
+    "turbodiffusion": TurboDiffusionT2V_1_3B_Config,
     # Other fallbacks by architecture
 }
 

fastvideo/configs/pipelines/turbodiffusion.py

@@ -0,0 +1,128 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+TurboDiffusion pipeline configurations.
+
+TurboDiffusion uses RCM (recurrent Consistency Model) scheduler with
+SLA (Sparse-Linear Attention) for fast 1-4 step video generation.
+"""
+from dataclasses import dataclass, field
+
+from fastvideo.configs.models import DiTConfig, EncoderConfig, VAEConfig
+from fastvideo.configs.models.dits import WanVideoConfig
+from fastvideo.configs.models.encoders import CLIPVisionConfig
+from fastvideo.configs.models.vaes import WanVAEConfig
+from fastvideo.configs.pipelines.base import PipelineConfig
+from fastvideo.configs.pipelines.wan import t5_postprocess_text, T5Config, BaseEncoderOutput
+
+import torch
+from collections.abc import Callable
+
+
+@dataclass
+class TurboDiffusionT2VConfig(PipelineConfig):
+    """Base configuration for TurboDiffusion T2V pipeline.
+    
+    Uses RCM scheduler with sigma_max=80 for 1-4 step generation.
+    No boundary_ratio (single model, no switching).
+    """
+    # DiT
+    dit_config: DiTConfig = field(default_factory=WanVideoConfig)
+    # VAE
+    vae_config: VAEConfig = field(default_factory=WanVAEConfig)
+    vae_tiling: bool = False
+    vae_sp: bool = False
+
+    # Denoising stage
+    flow_shift: float | None = 3.0
+
+    # No boundary_ratio for T2V (single model)
+    boundary_ratio: float | None = None
+
+    # Text encoding stage
+    text_encoder_configs: tuple[EncoderConfig, ...] = field(
+        default_factory=lambda: (T5Config(), ))
+    postprocess_text_funcs: tuple[Callable[[BaseEncoderOutput], torch.Tensor],
+                                  ...] = field(default_factory=lambda:
+                                               (t5_postprocess_text, ))
+
+    # Precision for each component
+    precision: str = "bf16"
+    vae_precision: str = "fp32"
+    text_encoder_precisions: tuple[str, ...] = field(
+        default_factory=lambda: ("fp32", ))
+
+    # self-forcing params
+    warp_denoising_step: bool = True
+
+    def __post_init__(self):
+        self.vae_config.load_encoder = False
+        self.vae_config.load_decoder = True
+        # Ensure no boundary_ratio is set in dit_config
+        self.dit_config.boundary_ratio = None
+
+
+@dataclass
+class TurboDiffusionT2V_1_3B_Config(TurboDiffusionT2VConfig):
+    """Configuration for TurboDiffusion T2V 1.3B model."""
+    pass
+
+
+@dataclass
+class TurboDiffusionT2V_14B_Config(TurboDiffusionT2VConfig):
+    """Configuration for TurboDiffusion T2V 14B model.
+    
+    Uses same config as 1.3B but with higher flow_shift for 14B model.
+    """
+    flow_shift: float | None = 5.0
+
+
+@dataclass
+class TurboDiffusionI2VConfig(PipelineConfig):
+    """Base configuration for TurboDiffusion I2V pipeline.
+    
+    Uses RCM scheduler with sigma_max=200 for 1-4 step generation.
+    Uses boundary_ratio=0.9 for high-noise to low-noise model switching.
+    """
+    # DiT
+    dit_config: DiTConfig = field(default_factory=WanVideoConfig)
+    # VAE
+    vae_config: VAEConfig = field(default_factory=WanVAEConfig)
+    vae_tiling: bool = False
+    vae_sp: bool = False
+
+    # Denoising stage
+    flow_shift: float | None = 5.0
+
+    boundary_ratio: float | None = 0.9
+
+    # Text encoding stage
+    text_encoder_configs: tuple[EncoderConfig, ...] = field(
+        default_factory=lambda: (T5Config(), ))
+    postprocess_text_funcs: tuple[Callable[[BaseEncoderOutput], torch.Tensor],
+                                  ...] = field(default_factory=lambda:
+                                               (t5_postprocess_text, ))
+
+    # Image encoder for I2V
+    image_encoder_config: EncoderConfig = field(
+        default_factory=CLIPVisionConfig)
+    image_encoder_precision: str = "fp32"
+
+    # Precision for each component
+    precision: str = "bf16"
+    vae_precision: str = "fp32"
+    text_encoder_precisions: tuple[str, ...] = field(
+        default_factory=lambda: ("fp32", ))
+
+    # self-forcing params
+    warp_denoising_step: bool = True
+
+    def __post_init__(self):
+        self.vae_config.load_encoder = True
+        self.vae_config.load_decoder = True
+        self.dit_config.boundary_ratio = self.boundary_ratio
+
+
+@dataclass
+class TurboDiffusionI2V_A14B_Config(TurboDiffusionI2VConfig):
+    """Configuration for TurboDiffusion I2V A14B model."""
+    pass

fastvideo/configs/sample/base.py

@@ -223,7 +223,7 @@ class SamplingParam:
             help="Path to input image for image-to-video generation",
         )
         parser.add_argument(
-            "--video_path",
+            "--video-path",
             type=str,
             default=SamplingParam.video_path,
             help="Path to input video for video-to-video generation",

fastvideo/configs/sample/cosmos2_5.py

@@ -0,0 +1,19 @@
+# SPDX-License-Identifier: Apache-2.0
+from dataclasses import dataclass
+
+from fastvideo.configs.sample.base import SamplingParam
+
+
+@dataclass
+class Cosmos_Predict2_5_2B_Diffusers_SamplingParam(SamplingParam):
+    """Defaults for Cosmos 2.5 (Predict2.5) text-to-video diffusers-format model."""
+
+    height: int = 480
+    width: int = 832
+    num_frames: int = 121
+    fps: int = 24
+
+    guidance_scale: float = 7.0
+    # Official Cosmos2.5 sampling uses empty string as unconditional.
+    negative_prompt: str = ""
+    num_inference_steps: int = 35

fastvideo/configs/sample/registry.py

@@ -9,6 +9,7 @@ from fastvideo.configs.sample.hunyuan15 import Hunyuan15_480P_SamplingParam, Hun
 from fastvideo.configs.sample.stepvideo import StepVideoT2VSamplingParam
 
 from fastvideo.configs.sample.cosmos import Cosmos_Predict2_2B_Video2World_SamplingParam
+from fastvideo.configs.sample.cosmos2_5 import Cosmos_Predict2_5_2B_Diffusers_SamplingParam
 
 # isort: off
 from fastvideo.configs.sample.wan import (
@@ -26,6 +27,11 @@ from fastvideo.configs.sample.wan import (
     SelfForcingWan2_2_T2V_A14B_480P_SamplingParam,
     MatrixGame2_SamplingParam,
 )
+from fastvideo.configs.sample.turbodiffusion import (
+    TurboDiffusionT2V_1_3B_SamplingParam,
+    TurboDiffusionT2V_14B_SamplingParam,
+    TurboDiffusionI2V_A14B_SamplingParam,
+)
 # isort: on
 from fastvideo.logger import init_logger
 from fastvideo.utils import (maybe_download_model_index,
@@ -34,36 +40,48 @@ from fastvideo.utils import (maybe_download_model_index,
 logger = init_logger(__name__)
 # Registry maps specific model weights to their config classes
 SAMPLING_PARAM_REGISTRY: dict[str, Any] = {
-    "FastVideo/FastHunyuan-diffusers": FastHunyuanSamplingParam,
-    "hunyuanvideo-community/HunyuanVideo": HunyuanSamplingParam,
+    "FastVideo/FastHunyuan-diffusers":
+    FastHunyuanSamplingParam,
+    "hunyuanvideo-community/HunyuanVideo":
+    HunyuanSamplingParam,
     "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-480p_t2v":
     Hunyuan15_480P_SamplingParam,
     "hunyuanvideo-community/HunyuanVideo-1.5-Diffusers-720p_t2v":
     Hunyuan15_720P_SamplingParam,
-    "FastVideo/stepvideo-t2v-diffusers": StepVideoT2VSamplingParam,
+    "FastVideo/stepvideo-t2v-diffusers":
+    StepVideoT2VSamplingParam,
 
     # Wan2.1
-    "Wan-AI/Wan2.1-T2V-1.3B-Diffusers": WanT2V_1_3B_SamplingParam,
-    "Wan-AI/Wan2.1-T2V-14B-Diffusers": WanT2V_14B_SamplingParam,
-    "Wan-AI/Wan2.1-I2V-14B-480P-Diffusers": WanI2V_14B_480P_SamplingParam,
-    "Wan-AI/Wan2.1-I2V-14B-720P-Diffusers": WanI2V_14B_720P_SamplingParam,
+    "Wan-AI/Wan2.1-T2V-1.3B-Diffusers":
+    WanT2V_1_3B_SamplingParam,
+    "Wan-AI/Wan2.1-T2V-14B-Diffusers":
+    WanT2V_14B_SamplingParam,
+    "Wan-AI/Wan2.1-I2V-14B-480P-Diffusers":
+    WanI2V_14B_480P_SamplingParam,
+    "Wan-AI/Wan2.1-I2V-14B-720P-Diffusers":
+    WanI2V_14B_720P_SamplingParam,
     "weizhou03/Wan2.1-Fun-1.3B-InP-Diffusers":
     Wan2_1_Fun_1_3B_InP_SamplingParam,
     "IRMChen/Wan2.1-Fun-1.3B-Control-Diffusers":
     Wan2_1_Fun_1_3B_Control_SamplingParam,
 
     # Wan2.2
-    "Wan-AI/Wan2.2-TI2V-5B-Diffusers": Wan2_2_TI2V_5B_SamplingParam,
+    "Wan-AI/Wan2.2-TI2V-5B-Diffusers":
+    Wan2_2_TI2V_5B_SamplingParam,
     "FastVideo/FastWan2.2-TI2V-5B-FullAttn-Diffusers":
     Wan2_2_TI2V_5B_SamplingParam,
-    "Wan-AI/Wan2.2-T2V-A14B-Diffusers": Wan2_2_T2V_A14B_SamplingParam,
-    "Wan-AI/Wan2.2-I2V-A14B-Diffusers": Wan2_2_I2V_A14B_SamplingParam,
+    "Wan-AI/Wan2.2-T2V-A14B-Diffusers":
+    Wan2_2_T2V_A14B_SamplingParam,
+    "Wan-AI/Wan2.2-I2V-A14B-Diffusers":
+    Wan2_2_I2V_A14B_SamplingParam,
 
     # FastWan2.1
-    "FastVideo/FastWan2.1-T2V-1.3B-Diffusers": FastWanT2V480P_SamplingParam,
+    "FastVideo/FastWan2.1-T2V-1.3B-Diffusers":
+    FastWanT2V480P_SamplingParam,
 
     # FastWan2.2
-    "FastVideo/FastWan2.2-TI2V-5B-Diffusers": Wan2_2_TI2V_5B_SamplingParam,
+    "FastVideo/FastWan2.2-TI2V-5B-Diffusers":
+    Wan2_2_TI2V_5B_SamplingParam,
 
     # Causal Self-Forcing Wan2.1
     "wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers":
@@ -79,10 +97,25 @@ SAMPLING_PARAM_REGISTRY: dict[str, Any] = {
     "nvidia/Cosmos-Predict2-2B-Video2World":
     Cosmos_Predict2_2B_Video2World_SamplingParam,
 
+    # Cosmos2.5
+    "KyleShao/Cosmos-Predict2.5-2B-Diffusers":
+    Cosmos_Predict2_5_2B_Diffusers_SamplingParam,
+
     # MatrixGame2.0 models
-    "FastVideo/Matrix-Game-2.0-Base-Diffusers": MatrixGame2_SamplingParam,
-    "FastVideo/Matrix-Game-2.0-GTA-Diffusers": MatrixGame2_SamplingParam,
-    "FastVideo/Matrix-Game-2.0-TempleRun-Diffusers": MatrixGame2_SamplingParam,
+    "FastVideo/Matrix-Game-2.0-Base-Diffusers":
+    MatrixGame2_SamplingParam,
+    "FastVideo/Matrix-Game-2.0-GTA-Diffusers":
+    MatrixGame2_SamplingParam,
+    "FastVideo/Matrix-Game-2.0-TempleRun-Diffusers":
+    MatrixGame2_SamplingParam,
+
+    # TurboDiffusion models
+    "loayrashid/TurboWan2.1-T2V-1.3B-Diffusers":
+    TurboDiffusionT2V_1_3B_SamplingParam,
+    "loayrashid/TurboWan2.1-T2V-14B-Diffusers":
+    TurboDiffusionT2V_14B_SamplingParam,
+    "loayrashid/TurboWan2.2-I2V-A14B-Diffusers":
+    TurboDiffusionI2V_A14B_SamplingParam,
 
     # Add other specific weight variants
 }
@@ -105,6 +138,12 @@ SAMPLING_PARAM_DETECTOR: dict[str, Callable[[str], bool]] = {
     lambda id: "wancausaldmdpipeline" in id.lower(),
     "matrixgame":
     lambda id: "matrixgame" in id.lower() or "matrix-game" in id.lower(),
+    "turbodiffusion":
+    lambda id: "turbodiffusion" in id.lower() or "turbowan" in id.lower(),
+    "cosmos25":
+    lambda id: "cosmos2_5" in id.lower(),
+    "cosmos":
+    lambda id: "cosmos" in id.lower() and "2_5" not in id.lower(),
     # Add other pipeline architecture detectors
 }
 
@@ -121,6 +160,10 @@ SAMPLING_FALLBACK_PARAM: dict[str, Any] = {
     "wancausaldmdpipeline": SelfForcingWan2_1_T2V_1_3B_480P_SamplingParam,
     "stepvideo": StepVideoT2VSamplingParam,
     "matrixgame": MatrixGame2_SamplingParam,
+    "turbodiffusion":
+    TurboDiffusionT2V_1_3B_SamplingParam,  # Default to T2V for fallback
+    "cosmos25": Cosmos_Predict2_5_2B_Diffusers_SamplingParam,
+    "cosmos": Cosmos_Predict2_2B_Video2World_SamplingParam,
     # Other fallbacks by architecture
 }
 
@@ -144,9 +187,6 @@ def get_sampling_param_cls_for_name(pipeline_name_or_path: str) -> Any | None:
 
     if os.path.exists(pipeline_name_or_path):
         config = verify_model_config_and_directory(pipeline_name_or_path)
-        logger.warning(
-            "FastVideo may not correctly identify the optimal sampling param for this model, as the local directory may have been renamed."
-        )
     else:
         config = maybe_download_model_index(pipeline_name_or_path)
 

fastvideo/configs/sample/turbodiffusion.py

@@ -0,0 +1,73 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+TurboDiffusion sampling parameters.
+
+TurboDiffusion uses RCM (recurrent Consistency Model) scheduler for
+1-4 step video generation with no classifier-free guidance.
+"""
+from dataclasses import dataclass
+
+from fastvideo.configs.sample.base import SamplingParam
+
+
+@dataclass
+class TurboDiffusionT2V_1_3B_SamplingParam(SamplingParam):
+    """Sampling parameters for TurboDiffusion T2V 1.3B model.
+    
+    Uses 4-step RCM sampling with guidance_scale=1.0 (no CFG).
+    """
+    # Video parameters
+    height: int = 480
+    width: int = 832
+    num_frames: int = 81
+    fps: int = 16
+
+    # Denoising stage - TurboDiffusion uses 1-4 steps with no CFG
+    guidance_scale: float = 1.0
+    num_inference_steps: int = 4
+
+    # No negative prompt needed for TurboDiffusion (no CFG)
+    negative_prompt: str | None = None
+
+
+@dataclass
+class TurboDiffusionT2V_14B_SamplingParam(SamplingParam):
+    """Sampling parameters for TurboDiffusion T2V 14B model.
+    
+    Uses 4-step RCM sampling with guidance_scale=1.0 (no CFG).
+    """
+    # Video parameters (720p for 14B)
+    height: int = 720
+    width: int = 1280
+    num_frames: int = 81
+    fps: int = 16
+
+    # Denoising stage - TurboDiffusion uses 1-4 steps with no CFG
+    guidance_scale: float = 1.0
+    num_inference_steps: int = 4
+
+    # No negative prompt needed for TurboDiffusion (no CFG)
+    negative_prompt: str | None = None
+
+
+@dataclass
+class TurboDiffusionI2V_A14B_SamplingParam(SamplingParam):
+    """Sampling parameters for TurboDiffusion I2V A14B model.
+    
+    Uses 4-step RCM sampling with dual-model switching (high/low noise).
+    """
+    # Video parameters (720p for A14B I2V)
+    height: int = 720
+    width: int = 1280
+    num_frames: int = 81
+    fps: int = 16
+
+    # Denoising stage - TurboDiffusion uses 1-4 steps with no CFG
+    guidance_scale: float = 1.0
+    num_inference_steps: int = 4
+
+    # Note: boundary_ratio is set in the pipeline config (TurboDiffusionI2VConfig),
+    # not here. This keeps sampling params and pipeline config separate.
+
+    # No negative prompt needed for TurboDiffusion (no CFG)
+    negative_prompt: str | None = None

fastvideo/dataset/__init__.py

@@ -8,6 +8,7 @@ from fastvideo.dataset.preprocessing_datasets import VideoCaptionMergedDataset,
 from fastvideo.dataset.transform import (CenterCropResizeVideo, Normalize255,
                                          TemporalRandomCrop)
 from fastvideo.dataset.validation_dataset import ValidationDataset
+from fastvideo.dataset.rl_prompt_dataset import build_rl_prompt_dataloader
 
 
 def getdataset(args) -> VideoCaptionMergedDataset:
@@ -47,5 +48,6 @@ def gettextdataset(args) -> TextDataset:
 
 __all__ = [
     "build_parquet_map_style_dataloader", "ValidationDataset",
-    "VideoCaptionMergedDataset", "TextDataset"
+    "VideoCaptionMergedDataset", "TextDataset",
+    "build_rl_prompt_dataloader"
 ]

fastvideo/dataset/rl_prompt_dataset.py

@@ -0,0 +1,174 @@
+# SPDX-License-Identifier: Apache-2.0
+import torch
+from torch.utils.data import Dataset, DataLoader, Sampler
+import json
+import os
+
+
+class TextPromptDataset(Dataset):
+    """Dataset for loading text prompts from a simple text file (one prompt per line)."""
+    
+    def __init__(self, dataset, split='train'):
+        self.file_path = os.path.join(dataset, f'{split}.txt')
+        with open(self.file_path, 'r') as f:
+            self.prompts = [line.strip() for line in f.readlines()]
+        
+    def __len__(self):
+        return len(self.prompts)
+    
+    def __getitem__(self, idx):
+        return {"prompt": self.prompts[idx], "metadata": {}}
+
+    @staticmethod
+    def collate_fn(examples):
+        prompts = [example["prompt"] for example in examples]
+        metadatas = [example["metadata"] for example in examples]
+        return prompts, metadatas
+
+
+class GenevalPromptDataset(Dataset):
+    """Dataset for loading prompts with metadata from JSONL files (e.g., GenEval format)."""
+    
+    def __init__(self, dataset, split='train'):
+        self.file_path = os.path.join(dataset, f'{split}_metadata.jsonl')
+        with open(self.file_path, 'r', encoding='utf-8') as f:
+            self.metadatas = [json.loads(line) for line in f]
+            self.prompts = [item['prompt'] for item in self.metadatas]
+        
+    def __len__(self):
+        return len(self.prompts)
+    
+    def __getitem__(self, idx):
+        return {"prompt": self.prompts[idx], "metadata": self.metadatas[idx]}
+
+    @staticmethod
+    def collate_fn(examples):
+        prompts = [example["prompt"] for example in examples]
+        metadatas = [example["metadata"] for example in examples]
+        return prompts, metadatas
+
+
+class KRepeatSampler(Sampler):
+    """Sampler that repeats each sample k times, ensuring synchronized random selection. For single-node training, set num_replicas=1 and rank=0."""
+    
+    def __init__(self, dataset, batch_size, k, num_replicas, rank, seed=0):
+        self.dataset = dataset
+        self.batch_size = batch_size        # Batch size per GPU/card
+        self.k = k                          # Number of repetitions per sample
+        self.num_replicas = num_replicas    # Total number of GPUs/cards
+        self.rank = rank                    # Current GPU/card rank
+        self.seed = seed                    # Random seed for synchronization
+        
+        # Calculate the number of unique samples needed for each iteration
+        self.total_samples = self.num_replicas * self.batch_size
+        assert self.total_samples % self.k == 0, f"k can not div n*b, k{k}-num_replicas{num_replicas}-batch_size{batch_size}"
+        self.m = self.total_samples // self.k  # different number of samples
+        self.step = 0
+
+    def __iter__(self):
+        while True:
+            # Generate a deterministic random sequence to ensure all cards are synchronized
+            g = torch.Generator()
+            g.manual_seed(self.seed + self.step)
+            
+            # Randomly select m unique samples
+            indices = torch.randperm(len(self.dataset), generator=g)[:self.m].tolist()
+            
+            # Repeat each sample k times to generate a total of n*b samples
+            repeated_indices = [idx for idx in indices for _ in range(self.k)]
+            
+            # Shuffle the order to ensure even distribution
+            shuffled_indices = torch.randperm(len(repeated_indices), generator=g).tolist()
+            shuffled_samples = [repeated_indices[i] for i in shuffled_indices]
+            
+            # Split samples among all cards
+            per_card_samples = []
+            for i in range(self.num_replicas):
+                start = i * self.batch_size
+                end = start + self.batch_size
+                per_card_samples.append(shuffled_samples[start:end])
+            
+            # Return the sample indices for the current card
+            yield per_card_samples[self.rank]
+        
+    def __len__(self):
+        return len(self.dataset) // self.batch_size
+    
+    def set_step(self, step):
+        """Used to synchronize the random state for different epochs."""
+        self.step = step
+
+
+def build_rl_prompt_dataloader(
+    dataset_path: str,
+    dataset_type: str = "text",
+    split: str = "train",
+    train_batch_size: int = 8,
+    test_batch_size: int = 8,
+    k: int = 1,
+    seed: int = 42,
+    train_num_workers: int = 1,
+    test_num_workers: int = 8,
+    num_replicas: int = 1,
+    rank: int = 0,
+) -> tuple[DataLoader, DataLoader]:
+    """
+    Factory function to create train and test dataloaders for RL prompt datasets.
+    
+    Args:
+        dataset_path: Path to dataset directory
+        dataset_type: "text" for TextPromptDataset or "geneval" for GenevalPromptDataset
+        split: Dataset split ("train" or "test")
+        train_batch_size: Batch size per GPU for training
+        test_batch_size: Batch size for testing
+        k: Number of times to repeat each sample (num_image_per_prompt)
+        seed: Random seed for sampler synchronization
+        train_num_workers: Number of workers for training dataloader
+        test_num_workers: Number of workers for test dataloader
+        num_replicas: Number of replicas (default 1 for single-node)
+        rank: Rank of current process (default 0 for single-node)
+    
+    Returns:
+        Tuple of (train_dataloader, test_dataloader)
+    """
+    # Create datasets based on type
+    if dataset_type == "text":
+        train_dataset = TextPromptDataset(dataset_path, 'train')
+        test_dataset = TextPromptDataset(dataset_path, 'test')
+        collate_fn = TextPromptDataset.collate_fn
+    elif dataset_type == "geneval":
+        train_dataset = GenevalPromptDataset(dataset_path, 'train')
+        test_dataset = GenevalPromptDataset(dataset_path, 'test')
+        collate_fn = GenevalPromptDataset.collate_fn
+    else:
+        raise ValueError(f"Unknown dataset_type: {dataset_type}. Must be 'text' or 'geneval'")
+    
+    # Create infinite-loop training sampler
+    train_sampler = KRepeatSampler(
+        dataset=train_dataset,
+        batch_size=train_batch_size,
+        k=k,
+        num_replicas=num_replicas,
+        rank=rank,
+        seed=seed
+    )
+    
+    # Create training dataloader with batch_sampler (infinite loop)
+    train_dataloader = DataLoader(
+        train_dataset,
+        batch_sampler=train_sampler,
+        num_workers=train_num_workers,
+        collate_fn=collate_fn,
+    )
+    
+    # Create standard test dataloader
+    test_dataloader = DataLoader(
+        test_dataset,
+        batch_size=test_batch_size,
+        collate_fn=collate_fn,
+        shuffle=False,
+        num_workers=test_num_workers,
+    )
+    
+    return train_dataloader, test_dataloader, train_dataset, test_dataset
+

fastvideo/fastvideo_args.py

@@ -132,8 +132,8 @@ class FastVideoArgs:
 
     # CPU offload parameters
     dit_cpu_offload: bool = True
-    use_fsdp_inference: bool = True
-    dit_layerwise_offload: bool = False
+    use_fsdp_inference: bool = False
+    dit_layerwise_offload: bool = True
     text_encoder_cpu_offload: bool = True
     image_encoder_cpu_offload: bool = True
     vae_cpu_offload: bool = True
@@ -431,7 +431,9 @@ class FastVideoArgs:
             "--use-fsdp-inference",
             action=StoreBoolean,
             help=
-            "Use FSDP for inference by sharding the model weights. Latency is very low due to prefetch--enable if run out of memory.",
+            "Use FSDP for inference by sharding the model weights. FSDP helps reduce GPU memory usage but may introduce"
+            +
+            " weight transfer overhead depending on the specific setup. Enable if run out of memory.",
         )
         parser.add_argument(
             "--text-encoder-cpu-offload",
@@ -738,6 +740,271 @@ def get_current_fastvideo_args() -> FastVideoArgs:
     return _current_fastvideo_args
 
 
+@dataclasses.dataclass
+class RLArgs:
+    """
+    Reinforcement Learning (RL) specific arguments
+    """
+    # ============================================================================
+    # SHARED RL CONFIGURATION
+    rl_mode: bool = False # Enable RL training mode
+    rl_algorithm: str = "grpo"  # RL algorithm to use: "grpo", "ppo", "dpo"
+
+    # Trajectory collection
+    num_rollouts: int = 4  # Number of rollouts to collect per training step
+    rollout_steps: str = "20,30"  # Random intermediate steps for sampling (comma-separated)
+    noise_injection_min: int = 10  # Minimum timestep for noise injection
+    noise_injection_max: int = 40  # Maximum timestep for noise injection
+    use_sde_sampling: bool = True  # Use SDE sampling (Flow-GRPO-Fast)
+    num_denoising_steps: int = 2  # Number of denoising steps per trajectory (1-2 for fast)
+
+    # Advantage estimation
+    gamma: float = 0.99  # Discount factor for returns
+    lambda_param: float = 0.95  # GAE lambda parameter
+    use_gae: bool = True  # Use Generalized Advantage Estimation
+    normalize_advantages: bool = True  # Normalize advantages before policy update
+
+    # Reward models
+    reward_models: dict[str, float] = field(default_factory=lambda: {"dummy": 1.0})  # reward models (names, weight)
+    value_model_path: str = ""  # Path to value model (can be empty to train from scratch)
+    value_model_share_backbone: bool = False  # Share transformer backbone between policy and value
+
+    # Training schedule
+    warmup_steps: int = 1000  # Collect SFT-style data before starting RL
+    collect_on_policy: bool = True  # Collect fresh rollouts each step (on-policy)
+    timestep_fraction: float = 0.99  # Fraction of timesteps to train on
+    num_inner_epochs: int = 1  # Number of inner epochs per outer epoch
+
+    # KL regularization
+    kl_beta: float = 0.004  # KL loss coefficient (GRPO uses KL loss, DPO uses larger beta)
+    kl_reward: float = 0.0  # KL reward coefficient (alternative to KL loss, typically 0)
+
+    # SFT integration
+    sft_weight: float = 0.0  # SFT loss weight for supervised learning in RL training
+    sft_batch_size: int = 3  # Batch size for SFT data
+
+    # CFG
+    guidance_scale = 1.0 # use guidance_scale > 1.0 to enable CFG
+
+    # Statistics tracking
+    global_std: bool = False  # Use global std across all samples vs per-group std
+    per_prompt_stat_tracking: bool = True  # Track statistics per prompt
+
+    # Training options
+    use_diffusion_loss: bool = True  # Use diffusion loss in training
+
+    # ============================================================================
+    # GRPO-SPECIFIC CONFIGURATION
+
+    # Policy optimization
+    grpo_policy_clip_range: float = 0.001  # PPO-style clipping range for policy ratio
+    grpo_value_clip_range: float = 0.2  # Value function clipping range
+    grpo_num_policy_epochs: int = 1  # Number of policy update epochs (GRPO typically uses 1)
+    grpo_num_value_epochs: int = 1  # Number of value function update epochs
+    grpo_target_kl: float = 0.01  # Target KL divergence for early stopping
+    grpo_entropy_coef: float = 0.0  # Entropy coefficient for exploration
+    grpo_value_loss_coef: float = 0.5  # Value loss coefficient
+
+    # GRPO-Guard safety mechanisms
+    grpo_use_grpo_guard: bool = True  # Enable GRPO-Guard safety mechanisms
+    grpo_ratio_norm_correction: bool = True  # RatioNorm: correct importance ratio bias
+    grpo_gradient_reweighting: bool = True  # Reweight gradients across denoising steps
+    grpo_max_importance_ratio: float = 10.0  # Clip importance ratios above this value
+
+    # ============================================================================
+    # DPO-SPECIFIC CONFIGURATION
+
+    dpo_beta: float = 100.0  # DPO regularization parameter (typically much larger than GRPO beta)
+    dpo_ref_update_step: int = 10000000  # Reference model update frequency for OnlineDPO
+    dpo_label_smoothing: float = 0.0  # Label smoothing for DPO loss
+
+    @staticmethod
+    def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
+        """Add RL-specific CLI arguments to the parser."""
+        # RL (Reinforcement Learning) arguments
+        parser.add_argument("--rl-mode",
+                            action=StoreBoolean,
+                            help="Enable RL training mode")
+        parser.add_argument("--rl-algorithm",
+                            type=str,
+                            default=RLArgs.rl_algorithm,
+                            choices=["grpo", "ppo", "dpo"],
+                            help="RL algorithm to use (grpo, ppo, dpo)")
+
+        # Trajectory collection (Flow-GRPO-Fast)
+        parser.add_argument("--rl-num-rollouts",
+                            type=int,
+                            default=RLArgs.num_rollouts,
+                            help="Number of rollouts to collect per training step")
+        parser.add_argument("--rl-rollout-steps",
+                            type=str,
+                            default=RLArgs.rollout_steps,
+                            help="Random intermediate steps for sampling (comma-separated)")
+        parser.add_argument("--rl-noise-injection-min",
+                            type=int,
+                            default=RLArgs.noise_injection_min,
+                            help="Minimum timestep for noise injection")
+        parser.add_argument("--rl-noise-injection-max",
+                            type=int,
+                            default=RLArgs.noise_injection_max,
+                            help="Maximum timestep for noise injection")
+        parser.add_argument("--rl-use-sde-sampling",
+                            action=StoreBoolean,
+                            help="Use SDE sampling (Flow-GRPO-Fast)")
+        parser.add_argument("--rl-num-denoising-steps",
+                            type=int,
+                            default=RLArgs.num_denoising_steps,
+                            help="Number of denoising steps per trajectory (1-2 for fast)")
+
+        # Advantage estimation
+        parser.add_argument("--rl-gamma",
+                            type=float,
+                            default=RLArgs.gamma,
+                            help="Discount factor for returns")
+        parser.add_argument("--rl-lambda",
+                            type=float,
+                            default=RLArgs.lambda_param,
+                            help="GAE lambda parameter")
+        parser.add_argument("--rl-use-gae",
+                            action=StoreBoolean,
+                            help="Use Generalized Advantage Estimation")
+        parser.add_argument("--rl-normalize-advantages",
+                            action=StoreBoolean,
+                            help="Normalize advantages before policy update")
+
+        # Policy optimization (GRPO/PPO)
+        parser.add_argument("--rl-policy-clip-range",
+                            type=float,
+                            default=RLArgs.grpo_policy_clip_range,
+                            dest="grpo_policy_clip_range",  # Map to RLArgs field name
+                            help="PPO-style clipping range for policy ratio")
+        parser.add_argument("--rl-value-clip-range",
+                            type=float,
+                            default=RLArgs.grpo_value_clip_range,
+                            help="Value function clipping range")
+        parser.add_argument("--rl-num-policy-epochs",
+                            type=int,
+                            default=RLArgs.grpo_num_policy_epochs,
+                            help="Number of policy update epochs (GRPO typically uses 1)")
+        parser.add_argument("--rl-num-value-epochs",
+                            type=int,
+                            default=RLArgs.grpo_num_value_epochs,
+                            help="Number of value function update epochs")
+        parser.add_argument("--rl-target-kl",
+                            type=float,
+                            default=RLArgs.grpo_target_kl,
+                            help="Target KL divergence for early stopping")
+        parser.add_argument("--rl-entropy-coef",
+                            type=float,
+                            default=RLArgs.grpo_entropy_coef,
+                            help="Entropy coefficient for exploration")
+        parser.add_argument("--rl-value-loss-coef",
+                            type=float,
+                            default=RLArgs.grpo_value_loss_coef,
+                            help="Value loss coefficient")
+
+        # GRPO-Guard (safety mechanisms)
+        parser.add_argument("--rl-use-grpo-guard",
+                            action=StoreBoolean,
+                            help="Enable GRPO-Guard safety mechanisms")
+        parser.add_argument("--rl-ratio-norm-correction",
+                            action=StoreBoolean,
+                            help="RatioNorm: correct importance ratio bias")
+        parser.add_argument("--rl-gradient-reweighting",
+                            action=StoreBoolean,
+                            help="Reweight gradients across denoising steps")
+        parser.add_argument("--rl-max-importance-ratio",
+                            type=float,
+                            default=RLArgs.grpo_max_importance_ratio,
+                            help="Clip importance ratios above this value")
+
+        # Reward models
+        parser.add_argument("--reward-models",
+                            type=str,
+                            default='{"dummy": 1.0}',
+                            help="Reward models as JSON dict (e.g., '{\"video_ocr\": 1.0, \"pickscore\": 0.5}')")
+        parser.add_argument("--value-model-path",
+                            type=str,
+                            default=RLArgs.value_model_path,
+                            help="Path to value model (can be empty to train from scratch)")
+        parser.add_argument("--value-model-share-backbone",
+                            action=StoreBoolean,
+                            help="Share transformer backbone between policy and value")
+
+        # Training schedule
+        parser.add_argument("--rl-warmup-steps",
+                            type=int,
+                            default=RLArgs.warmup_steps,
+                            help="Collect SFT-style data before starting RL")
+        parser.add_argument("--rl-collect-on-policy",
+                            action=StoreBoolean,
+                            help="Collect fresh rollouts each step (on-policy)")
+        parser.add_argument("--rl-timestep-fraction",
+                            type=float,
+                            default=RLArgs.timestep_fraction,
+                            help="Fraction of timesteps to train on")
+        parser.add_argument("--rl-num-inner-epochs",
+                            type=int,
+                            default=RLArgs.num_inner_epochs,
+                            help="Number of inner epochs per outer epoch")
+
+        # KL regularization
+        parser.add_argument("--rl-kl-beta",
+                            type=float,
+                            default=RLArgs.kl_beta,
+                            dest="kl_beta",  # Map CLI arg to RLArgs field name
+                            help="KL loss coefficient (GRPO uses KL loss, DPO uses larger beta)")
+        parser.add_argument("--rl-kl-reward",
+                            type=float,
+                            default=RLArgs.kl_reward,
+                            help="KL reward coefficient (alternative to KL loss, typically 0)")
+
+        # SFT integration
+        parser.add_argument("--rl-sft-weight",
+                            type=float,
+                            default=RLArgs.sft_weight,
+                            help="SFT loss weight for supervised learning in RL training")
+        parser.add_argument("--rl-sft-batch-size",
+                            type=int,
+                            default=RLArgs.sft_batch_size,
+                            help="Batch size for SFT data")
+
+        # CFG settings
+        parser.add_argument("--guidance-scale",
+                            type=float,
+                            default=1.0,
+                            help="Guidance scale for CFG")
+
+        # Statistics tracking
+        parser.add_argument("--rl-global-std",
+                            action=StoreBoolean,
+                            help="Use global std across all samples vs per-group std")
+        parser.add_argument("--rl-per-prompt-stat-tracking",
+                            action=StoreBoolean,
+                            help="Track statistics per prompt")
+
+        # Training options
+        parser.add_argument("--rl-use-diffusion-loss",
+                            action=StoreBoolean,
+                            help="Use diffusion loss in training")
+
+        # DPO-specific
+        parser.add_argument("--dpo-beta",
+                            type=float,
+                            default=RLArgs.dpo_beta,
+                            help="DPO regularization parameter (typically much larger than GRPO beta)")
+        parser.add_argument("--dpo-ref-update-step",
+                            type=int,
+                            default=RLArgs.dpo_ref_update_step,
+                            help="Reference model update frequency for OnlineDPO")
+        parser.add_argument("--dpo-label-smoothing",
+                            type=float,
+                            default=RLArgs.dpo_label_smoothing,
+                            help="Label smoothing for DPO loss")
+
+        return parser
+
+
 @dataclasses.dataclass
 class TrainingArgs(FastVideoArgs):
     """
@@ -750,6 +1017,11 @@ class TrainingArgs(FastVideoArgs):
     num_height: int = 0
     num_width: int = 0
     num_frames: int = 0
+    
+    # RL dataset configuration (for RL prompt datasets)
+    rl_dataset_path: str = ""  # Path to RL prompt dataset directory (defaults to data_path if not set)
+    rl_dataset_type: str = "text"  # "text" or "geneval"
+    rl_num_image_per_prompt: int = 4  # k parameter for KRepeatSampler (num_image_per_prompt)
 
     train_batch_size: int = 0
     num_latent_t: int = 0
@@ -860,6 +1132,9 @@ class TrainingArgs(FastVideoArgs):
     last_step_only: bool = False  # Only use the last timestep for training
     context_noise: int = 0  # Context noise level for cache updates
 
+    # Nested RL configuration
+    rl_args: RLArgs = dataclasses.field(default_factory=RLArgs)
+
     @classmethod
     def from_cli_args(cls, args: argparse.Namespace) -> "TrainingArgs":
         provided_args = clean_cli_args(args)
@@ -884,6 +1159,25 @@ class TrainingArgs(FastVideoArgs):
                 kwargs[attr] = WorkloadType.from_string(
                     workload_type_value) if isinstance(
                         workload_type_value, str) else workload_type_value
+            elif attr == 'rl_args':
+                # Construct nested RLArgs from CLI arguments
+                rl_kwargs = {}
+                for rl_field in dataclasses.fields(RLArgs):
+                    rl_attr = rl_field.name
+                    if hasattr(args, rl_attr):
+                        value = getattr(args, rl_attr)
+                        # Special handling for reward_models: parse JSON string to dict
+                        if rl_attr == 'reward_models' and isinstance(value, str):
+                            rl_kwargs[rl_attr] = json.loads(value) if value else {}
+                        else:
+                            rl_kwargs[rl_attr] = value
+                    else:
+                        # Use default value from RLArgs
+                        if rl_field.default_factory is not dataclasses.MISSING:
+                            rl_kwargs[rl_attr] = rl_field.default_factory()
+                        elif rl_field.default is not dataclasses.MISSING:
+                            rl_kwargs[rl_attr] = rl_field.default
+                kwargs[attr] = RLArgs(**rl_kwargs)
             # Use getattr with default value from the dataclass for potentially missing attributes
             else:
                 # Get the field to check its default value
@@ -913,11 +1207,26 @@ class TrainingArgs(FastVideoArgs):
         parser.add_argument("--data-path",
                             type=str,
                             required=True,
-                            help="Path to parquet files")
+                            help="Path to parquet files (or RL prompt dataset directory for RL training)")
         parser.add_argument("--dataloader-num-workers",
                             type=int,
                             required=True,
                             help="Number of workers for dataloader")
+        
+        # RL dataset arguments (optional, defaults to data_path)
+        parser.add_argument("--rl-dataset-path",
+                            type=str,
+                            default="",
+                            help="Path to RL prompt dataset directory (defaults to --data-path if not set)")
+        parser.add_argument("--rl-dataset-type",
+                            type=str,
+                            default="text",
+                            choices=["text", "geneval"],
+                            help="RL dataset type: 'text' for TextPromptDataset or 'geneval' for GenevalPromptDataset")
+        parser.add_argument("--rl-num-image-per-prompt",
+                            type=int,
+                            default=4,
+                            help="Number of times to repeat each prompt (k parameter for KRepeatSampler)")
         parser.add_argument("--num-height",
                             type=int,
                             required=True,
@@ -1282,6 +1591,9 @@ class TrainingArgs(FastVideoArgs):
                             default=TrainingArgs.context_noise,
                             help="Context noise level for cache updates")
 
+        # RL (Reinforcement Learning) arguments
+        RLArgs.add_cli_args(parser)
+
         return parser
 
 

fastvideo/hooks/hooks.py

@@ -0,0 +1,122 @@
+# SPDX-License-Identifier: Apache-2.0
+
+import functools
+from typing import Any
+from torch import nn
+
+
+class ForwardHook:
+    """
+    Base class for forward hooks.
+    Hooks are used in the way:
+        modified_args, modified_kwargs = hook.pre_forward(module, *args, **kwargs)
+        output = module.forward(*modified_args, **modified_kwargs)
+        modified_output = hook.post_forward(module, output)
+    """
+
+    @classmethod
+    def name(cls) -> str:
+        raise NotImplementedError
+
+    def on_attach(self, module: nn.Module):  # noqa: B027
+        """Called once when the hook is attached to the module."""
+        pass
+
+    def on_detach(self, module: nn.Module):  # noqa: B027
+        """
+        Called once when the hook is detached from the module.
+        Note: this function is not guaranteed to be called if the module is
+              deleted before the hook is detached.
+        """
+        pass
+
+    def pre_forward(self, module: nn.Module, *args,
+                    **kwargs) -> tuple[tuple[Any, ...], dict[str, Any]]:
+        """Called before the module's forward method is executed."""
+        return args, kwargs
+
+    def post_forward(self, module: nn.Module, output: Any) -> Any:
+        """Called after the module's forward method is executed."""
+        return output
+
+
+class ModuleHookManager:
+    module_hook_attribute = "_hook_manager"
+
+    def __init__(self, module: nn.Module):
+        self.module = module
+        self.forward_hooks: dict[str, ForwardHook] = {}
+        self.original_forward = module.forward
+
+    @classmethod
+    def get_from(cls, module: nn.Module) -> "ModuleHookManager | None":
+        if hasattr(module, cls.module_hook_attribute):
+            return getattr(module, cls.module_hook_attribute)
+        return None
+
+    @classmethod
+    def get_from_or_default(cls, module: nn.Module) -> "ModuleHookManager":
+        if not hasattr(module, cls.module_hook_attribute):
+            setattr(module, cls.module_hook_attribute, cls(module))
+
+            def forward_hook_wrapper(mod: nn.Module, *args, **kwargs):
+                manager: ModuleHookManager = getattr(mod,
+                                                     cls.module_hook_attribute)
+                for hook in manager.forward_hooks.values():
+                    args, kwargs = hook.pre_forward(mod, *args, **kwargs)
+                output = manager.original_forward(*args, **kwargs)
+                for hook in reversed(manager.forward_hooks.values()):
+                    output = hook.post_forward(mod, output)
+                return output
+
+            module.forward = functools.partial(forward_hook_wrapper, module)
+
+        return getattr(module, cls.module_hook_attribute)
+
+    @staticmethod
+    def remove_from_manager(module: nn.Module) -> None:
+        if hasattr(module, ModuleHookManager.module_hook_attribute):
+            manager: ModuleHookManager = getattr(
+                module, ModuleHookManager.module_hook_attribute)
+            module.forward = manager.original_forward
+            delattr(module, ModuleHookManager.module_hook_attribute)
+
+    def _check_manager_attached(self) -> None:
+        if not hasattr(self.module, self.module_hook_attribute):
+            raise ValueError("ModuleHookManager is not attached to the module.")
+        if getattr(self.module, self.module_hook_attribute) is not self:
+            raise ValueError(
+                "ModuleHookManager attached to the module is different.")
+
+    def append_forward_hook(self, hook: ForwardHook):
+        self._check_manager_attached()
+        if hook.name() in self.forward_hooks:
+            raise ValueError(
+                f"Hook with name {hook.name()} is already registered.")
+        # after python 3.7, dicts maintain insertion order
+        self.forward_hooks[hook.name()] = hook
+        hook.on_attach(self.module)
+
+    def replace_forward_hook(self,
+                             hook_name: str,
+                             new_hook: ForwardHook,
+                             run_on_attach: bool = True):
+        self._check_manager_attached()
+        if hook_name not in self.forward_hooks:
+            raise ValueError(f"No hook with name {hook_name} found.")
+        old_hook = self.forward_hooks[hook_name]
+        if run_on_attach:
+            old_hook.on_detach(self.module)
+        self.forward_hooks[hook_name] = new_hook
+        new_hook.on_attach(self.module)
+
+    def remove_forward_hook(self, hook_name: str, run_detach: bool = True):
+        self._check_manager_attached()
+        if hook_name not in self.forward_hooks:
+            raise ValueError(f"No hook with name {hook_name} found.")
+        if run_detach:
+            self.forward_hooks[hook_name].on_detach(self.module)
+        del self.forward_hooks[hook_name]
+
+    def get_forward_hook(self, hook_name: str) -> ForwardHook | None:
+        return self.forward_hooks.get(hook_name, None)

fastvideo/hooks/layerwise_offload.py

@@ -0,0 +1,164 @@
+from contextlib import contextmanager
+from typing import Any
+import torch
+from torch import nn
+from fastvideo.hooks.hooks import ForwardHook, ModuleHookManager
+from fastvideo.logger import init_logger
+
+logger = init_logger(__name__)
+
+
+def _tensor_placeholder(tensor: torch.Tensor,
+                        device: torch.device) -> torch.Tensor:
+    """Create a rank-preserving empty placeholder on the specified device."""
+    shape = (0, ) if tensor.ndim <= 0 else (0, ) * tensor.ndim
+    return torch.empty(shape, device=device, dtype=tensor.dtype)
+
+
+class LayerwiseOffloadState:
+
+    def __init__(
+        self,
+        async_copy_stream: torch.cuda.Stream,
+        device: torch.device,
+        next_state: "LayerwiseOffloadState | None" = None,
+    ) -> None:
+        self.async_copy_stream = async_copy_stream
+        self.next_state = next_state
+        self.gpu_named_parameters: dict[str, torch.Tensor] = {}
+        self.cpu_named_parameters: dict[str, torch.Tensor] = {}
+        self.module_ref: nn.Module = None  # type: ignore
+        self.device: torch.device = device
+
+    def _will_offload(self, name: str) -> bool:
+        return True
+
+    @torch.compiler.disable
+    def on_init(self, module: nn.Module):
+        self.module_ref = module
+        for name, param in self.module_ref.named_parameters():
+            if self._will_offload(name):
+                self.cpu_named_parameters[name] = (
+                    param.data.detach().to("cpu").pin_memory())
+                param.data = _tensor_placeholder(param.data, self.device)
+
+    @torch.compiler.disable
+    def wait_and_replace_params(self):
+        torch.cuda.current_stream().wait_stream(self.async_copy_stream)
+        # now gpu_named_parameters are ready
+        for name, param in self.module_ref.named_parameters():
+            if not self._will_offload(name):
+                continue
+            if name not in self.gpu_named_parameters:
+                # first load with blocking load
+                self.gpu_named_parameters[name] = self.cpu_named_parameters[
+                    name].to(self.device)
+            param.data = self.gpu_named_parameters[name]
+
+    @torch.compiler.disable
+    def prefetch_params(self):
+        compute_stream = torch.cuda.current_stream()
+        with torch.cuda.stream(self.async_copy_stream):
+            for name, param in self.module_ref.named_parameters():
+                if not self._will_offload(name):
+                    continue
+                assert name not in self.gpu_named_parameters
+                gpu_param = self.cpu_named_parameters[name].to(
+                    self.device, non_blocking=True)
+                gpu_param.record_stream(
+                    compute_stream
+                )  # ensure tensor will not be freed until forward is completed
+                self.gpu_named_parameters[name] = gpu_param
+
+    @torch.compiler.disable
+    def release_gpu_params(self):
+        for name, param in self.module_ref.named_parameters():
+            if self._will_offload(name):
+                param.data = _tensor_placeholder(param.data, self.device)
+                del self.gpu_named_parameters[name]
+        assert len(self.gpu_named_parameters) == 0
+
+
+class LayerwiseOffloadHook(ForwardHook):
+    """A hook that enables layerwise CPU offloading during forward pass."""
+
+    def __init__(self, state: LayerwiseOffloadState) -> None:
+        self.state = state
+
+    def on_attach(self, module: nn.Module):
+        self.state.on_init(module)  # pyright: ignore
+
+    def on_detach(self, module: nn.Module):
+        named_parameters = dict(module.named_parameters())
+        for name, cpu_tensor in self.state.cpu_named_parameters.items():
+            if name not in self.state.gpu_named_parameters:
+                if name in named_parameters:
+                    named_parameters[name].data = cpu_tensor.to(
+                        device=self.state.device)
+                else:
+                    logger.warning(
+                        "Parameter {} not found in module during detachment.",
+                        name,
+                    )
+
+    @classmethod
+    def name(cls) -> str:
+        return "LayerwiseOffloadHook"
+
+    def pre_forward(self, module: nn.Module, *args, **kwargs):
+        self.state.wait_and_replace_params()  # pyright: ignore
+        if self.state.next_state is not None:
+            self.state.next_state.prefetch_params()  # pyright: ignore
+        return args, kwargs
+
+    def post_forward(self, module: torch.nn.Module, output: Any):
+        self.state.release_gpu_params()  # pyright: ignore
+        return output
+
+    @contextmanager
+    def mutate_params_scope(self):
+        try:
+            # load params to GPU and keep them there
+            self.state.wait_and_replace_params()  # pyright: ignore
+            yield
+        finally:
+            # instead of releasing, we should overwrite the original params since they have been modified
+            self.state.cpu_named_parameters.clear()
+            self.state.gpu_named_parameters.clear()
+            self.state.on_init(self.state.module_ref)  # pyright: ignore
+
+
+def enable_layerwise_offload(model: nn.Module, is_replace: bool = False):
+    if torch.cuda.is_available():
+        device = torch.device("cuda", torch.cuda.current_device())
+    else:
+        logger.warning(
+            "CUDA is not available. Layerwise offloading is disabled.")
+        return
+    state_list = []
+    async_stream = torch.cuda.Stream()
+    for name, submodule in model.named_children():
+        if isinstance(submodule, nn.ModuleList):
+            for idx, module_entry in enumerate(submodule):
+                state = LayerwiseOffloadState(async_copy_stream=async_stream,
+                                              device=device)
+                state_list.append(state)
+                hook_mgr = ModuleHookManager.get_from_or_default(module_entry)
+                hook = LayerwiseOffloadHook(state)
+                if is_replace:
+                    existing_hook = hook_mgr.forward_hooks.get(hook.name())
+                    if existing_hook is not None:
+                        hook_mgr.replace_forward_hook(hook.name(), hook)
+                    else:
+                        raise AssertionError(
+                            f"Expect hook exists in {name} for replacement.")
+                else:
+                    hook_mgr.append_forward_hook(hook)
+            break
+    if len(state_list) == 0:
+        raise ValueError(
+            "No nn.ModuleList found in the model for layerwise offloading.")
+
+    # circular linking of states
+    for i in range(len(state_list)):
+        state_list[i].next_state = state_list[(i + 1) % len(state_list)]

fastvideo/models/dits/longcat.py

@@ -1,9 +1,6 @@
 # SPDX-License-Identifier: Apache-2.0
 """
 Native LongCat Video DiT implementation using FastVideo conventions.
-
-This is a Phase 2 reimplementation that replaces the third_party wrapper
-with native FastVideo layers for better performance and integration.
 """
 
 from typing import Any
@@ -129,7 +126,7 @@ class TimestepEmbedder(nn.Module):
         # Sinusoidal embedding in FP32
         t_freq = self.timestep_embedding(t.flatten(), self.frequency_embedding_size)
         
-        # Cast to model dtype before MLP
+        # Cast to model dtype before MLP (matching original LongCat)
         # Handle LoRA wrapper if present
         linear_layer = self.linear_1.base_layer if hasattr(self.linear_1, 'base_layer') else self.linear_1
         target_dtype = linear_layer.weight.dtype
@@ -166,13 +163,14 @@ class CaptionEmbedder(nn.Module):
         self.text_tokens_zero_pad = text_tokens_zero_pad
 
         # Two-layer MLP using ReplicatedLinear
+        # CRITICAL: Original LongCat uses GELU(approximate="tanh"), NOT SiLU!
         self.linear_1 = ReplicatedLinear(
             caption_channels,
             hidden_size,
             bias=True,
             params_dtype=dtype,
         )
-        self.act = nn.SiLU()
+        self.act = nn.GELU(approximate="tanh")  # Match original LongCat
         self.linear_2 = ReplicatedLinear(
             hidden_size,
             hidden_size,
@@ -268,10 +266,19 @@ class LongCatSelfAttention(nn.Module):
         self,
         x: torch.Tensor,              # [B, N, C]
         latent_shape: tuple,          # (T, H, W)
+        num_cond_latents: int = 0,    # Number of conditioning latent frames (for I2V)
+        return_kv: bool = False,      # Return K/V for caching
         **kwargs
-    ) -> torch.Tensor:
+    ) -> torch.Tensor | tuple:
         """
         Forward pass with 3D RoPE and optional BSA.
+        
+        For I2V mode (num_cond_latents > 0):
+        - Conditioned tokens only attend to themselves
+        - Noise tokens attend to ALL tokens (cond + noise)
+        
+        Args:
+            return_kv: If True, return (output, (k_cache, v_cache)) for KV caching
         """
         B, N, C = x.shape
         T, H, W = latent_shape
@@ -290,6 +297,12 @@ class LongCatSelfAttention(nn.Module):
         q = self.q_norm(q)
         k = self.k_norm(k)
 
+        # Save pre-RoPE K/V for cache if requested (before RoPE is applied)
+        if return_kv:
+            # [B, N, num_heads, head_dim] -> [B, num_heads, N, head_dim]
+            k_cache = k.transpose(1, 2).clone()
+            v_cache = v.transpose(1, 2).clone()
+
         # For RoPE: need [B, num_heads, N, head_dim]
         q_rope = q.transpose(1, 2)
         k_rope = k.transpose(1, 2)
@@ -301,6 +314,50 @@ class LongCatSelfAttention(nn.Module):
         q = q_rope.transpose(1, 2)
         k = k_rope.transpose(1, 2)
 
+        # === I2V Split Attention ===
+        # For I2V, conditioned tokens and noise tokens are processed separately
+        if num_cond_latents > 0:
+            # Calculate number of conditioned tokens (cond_latents * spatial_tokens_per_frame)
+            num_cond_tokens = num_cond_latents * (N // T)
+            
+            # Conditioned tokens: only attend to themselves (same seq length, use self.attn)
+            q_cond = q[:, :num_cond_tokens].contiguous()
+            k_cond = k[:, :num_cond_tokens].contiguous()
+            v_cond = v[:, :num_cond_tokens].contiguous()
+            out_cond, _ = self.attn(q_cond, k_cond, v_cond)
+            
+            # Noise tokens: attend to ALL tokens (different seq lengths!)
+            # Need to use flash attention directly since q has different length than k/v
+            q_noise = q[:, num_cond_tokens:].contiguous()  # [B, N_noise, num_heads, head_dim]
+            # k, v are full: [B, N, num_heads, head_dim]
+            
+            # Transpose for flash attention: [B, num_heads, seq, head_dim]
+            q_noise_t = q_noise.transpose(1, 2)
+            k_t = k.transpose(1, 2)
+            v_t = v.transpose(1, 2)
+            
+            # Use scaled dot product attention (handles different q/kv lengths)
+            out_noise_t = torch.nn.functional.scaled_dot_product_attention(
+                q_noise_t, k_t, v_t, 
+                attn_mask=None,
+                dropout_p=0.0,
+                is_causal=False
+            )  # [B, num_heads, N_noise, head_dim]
+            
+            # Transpose back: [B, N_noise, num_heads, head_dim]
+            out_noise = out_noise_t.transpose(1, 2)
+            
+            # Merge conditioned and noise outputs
+            out = torch.cat([out_cond, out_noise], dim=1)
+            
+            # Reshape and project out
+            out = out.reshape(B, N, C)
+            out, _ = self.to_out(out)
+            
+            if return_kv:
+                return out, (k_cache, v_cache)
+            return out
+
         # === Attention: BSA or standard ===
         if self.enable_bsa and T > 1:  # Only use BSA for multi-frame videos
             # BSA expects [B, H, S, D] format
@@ -348,6 +405,96 @@ class LongCatSelfAttention(nn.Module):
         out = out.reshape(B, N, C)
         out, _ = self.to_out(out)
 
+        if return_kv:
+            return out, (k_cache, v_cache)
+        return out
+
+    def forward_with_kv_cache(
+        self,
+        x: torch.Tensor,              # [B, N_noise, C] - only noise tokens
+        latent_shape: tuple,          # (T_noise, H, W) - shape for noise only
+        num_cond_latents: int,        # Number of conditioning latent frames
+        kv_cache: tuple,              # (k_cond, v_cond) - [B, heads, N_cond, head_dim]
+    ) -> torch.Tensor:
+        """
+        Forward using cached K/V from conditioning frames.
+        
+        x contains only NOISE tokens.
+        kv_cache contains pre-computed K/V for CONDITIONING tokens.
+        
+        CRITICAL: RoPE positions for noise tokens must start AFTER conditioning.
+        We achieve this by padding Q with dummy tokens for conditioning positions,
+        applying RoPE to the full sequence, then extracting only noise token Q.
+        """
+        B, N, C = x.shape
+        T, H, W = latent_shape
+
+        k_cache, v_cache = kv_cache
+
+        # Handle batch size mismatch (cache might be smaller for CFG)
+        # When using CFG, latent_model_input is doubled [neg, pos], but cache is for original batch
+        if k_cache.shape[0] != B:
+            # Expand cache to match input batch size
+            # For CFG: repeat the cache for both negative and positive branches
+            repeat_factor = B // k_cache.shape[0]
+            k_cache = k_cache.repeat(repeat_factor, 1, 1, 1)
+            v_cache = v_cache.repeat(repeat_factor, 1, 1, 1)
+
+        # Project to Q/K/V for noise tokens
+        q, _ = self.to_q(x)
+        k, _ = self.to_k(x)
+        v, _ = self.to_v(x)
+
+        # Reshape to heads: [B, N, num_heads, head_dim]
+        q = q.view(B, N, self.num_heads, self.head_dim)
+        k = k.view(B, N, self.num_heads, self.head_dim)
+        v = v.view(B, N, self.num_heads, self.head_dim)
+
+        # Per-head RMS normalization
+        q = self.q_norm(q)
+        k = self.k_norm(k)
+
+        # Transpose for RoPE: [B, heads, N, head_dim]
+        q_rope = q.transpose(1, 2)
+        k_rope = k.transpose(1, 2)
+        v = v.transpose(1, 2)
+
+        # CRITICAL: Apply RoPE with correct positional offset
+        # Noise frame queries need positions starting from num_cond_latents
+        # Following the original LongCat approach:
+        # 1. Pad Q with dummy tokens matching k_cache shape
+        # 2. Apply RoPE to full sequence (T_cond + T_noise)
+        # 3. Extract only the noise portion of Q
+        
+        # Create dummy Q padding to fill conditioning positions
+        # k_cache shape: [B, heads, N_cond, head_dim]
+        q_padding = torch.cat([torch.empty_like(k_cache), q_rope], dim=2).contiguous()
+        
+        # Concatenate cached K with noise K for RoPE
+        k_full = torch.cat([k_cache, k_rope], dim=2)
+        v_full = torch.cat([v_cache, v], dim=2)
+        
+        # Apply RoPE to full sequence (includes both cond and noise positions)
+        # Grid size: (T_cond + T_noise, H, W)
+        full_T = num_cond_latents + T
+        q_padding, k_full = self.rope_3d(q_padding, k_full, grid_size=(full_T, H, W))
+        
+        # Extract only the noise portion of Q (last N tokens)
+        q_rope = q_padding[:, :, -N:].contiguous()
+
+        # Run attention: Q_noise attends to full K/V (cond + noise)
+        out = torch.nn.functional.scaled_dot_product_attention(
+            q_rope, k_full, v_full,
+            attn_mask=None,
+            dropout_p=0.0,
+            is_causal=False
+        )  # [B, heads, N_noise, head_dim]
+
+        # Transpose back: [B, N_noise, heads, head_dim]
+        out = out.transpose(1, 2)
+        out = out.reshape(B, N, C)
+        out, _ = self.to_out(out)
+
         return out
 
 
@@ -394,6 +541,8 @@ class LongCatCrossAttention(nn.Module):
         self,
         x: torch.Tensor,              # [B, N_img, C]
         context: torch.Tensor,        # [B, N_text, C]
+        latent_shape: tuple = None,   # (T, H, W) - needed for I2V
+        num_cond_latents: int = 0,    # Number of conditioning latent frames (for I2V)
         **kwargs
     ) -> torch.Tensor:
         """
@@ -402,9 +551,57 @@ class LongCatCrossAttention(nn.Module):
         Args:
             x: Image tokens [B, N_img, C]
             context: Text tokens [B, N_text, C] (standard padded format)
+            latent_shape: (T, H, W) - needed for calculating num_cond_tokens
+            num_cond_latents: Number of conditioning latent frames (for I2V)
+            
+        For I2V mode (num_cond_latents > 0):
+        - Conditioned tokens get ZERO cross-attention output
+        - Only noise tokens get cross-attention with text
         """
         B, N_img, C = x.shape
 
+        # === I2V: Only noise tokens get cross-attention ===
+        if num_cond_latents > 0 and latent_shape is not None:
+            T, H, W = latent_shape
+            num_cond_tokens = num_cond_latents * (N_img // T)
+            
+            # Only process noise tokens
+            x_noise = x[:, num_cond_tokens:]  # [B, N_noise, C]
+            
+            # Project Q, K, V for noise tokens only
+            q, _ = self.to_q(x_noise)
+            k, _ = self.to_k(context)
+            v, _ = self.to_v(context)
+            
+            N_text = context.shape[1]
+            N_noise = x_noise.shape[1]
+
+            # Reshape to heads
+            q = q.view(B, N_noise, self.num_heads, self.head_dim)
+            k = k.view(B, N_text, self.num_heads, self.head_dim)
+            v = v.view(B, N_text, self.num_heads, self.head_dim)
+
+            # Per-head RMS normalization
+            q = self.q_norm(q)
+            k = self.k_norm(k)
+
+            # Run cross-attention
+            out_noise = self.attn(q, k, v)  # [B, N_noise, num_heads, head_dim]
+            out_noise = out_noise.reshape(B, N_noise, C)
+            out_noise, _ = self.to_out(out_noise)
+            
+            # Conditioned tokens get zero output
+            out_cond = torch.zeros(
+                (B, num_cond_tokens, C), 
+                dtype=out_noise.dtype, 
+                device=out_noise.device
+            )
+            
+            # Merge
+            out = torch.cat([out_cond, out_noise], dim=1)
+            return out
+
+        # === Standard cross-attention ===
         # Project Q, K, V (standard cross-attention like WanVideo/StepVideo/Cosmos)
         q, _ = self.to_q(x)
         k, _ = self.to_k(context)
@@ -475,19 +672,19 @@ class LongCatSwiGLUFFN(nn.Module):
 
 def modulate_fp32(norm: nn.Module, x: torch.Tensor, shift: torch.Tensor, scale: torch.Tensor) -> torch.Tensor:
     """
-    Apply modulation in FP32 for numerical stability (matching original LongCat).
+    Apply modulation in FP32 for numerical stability.
     
-    shift and scale should already be FP32 from torch.amp.autocast context.
+    Converts inputs to FP32 for the modulation operation, then casts back.
     """
-    # Ensure modulation params are FP32 (should be from autocast)
-    assert shift.dtype == torch.float32 and scale.dtype == torch.float32, \
-        f"shift and scale must be FP32, got {shift.dtype} and {scale.dtype}"
-    
     orig_dtype = x.dtype
     
+    # Convert to FP32 for numerical stability
+    shift_fp32 = shift.float()
+    scale_fp32 = scale.float()
+    
     # Normalize and modulate in FP32
     x_norm = norm(x.to(torch.float32))
-    x_mod = x_norm * (scale + 1) + shift
+    x_mod = x_norm * (scale_fp32 + 1) + shift_fp32
     
     return x_mod.to(orig_dtype)
 
@@ -568,10 +765,21 @@ class LongCatTransformerBlock(nn.Module):
         context: torch.Tensor,        # [B, N_text, C]
         t: torch.Tensor,              # [B, T, C_t]
         latent_shape: tuple,          # (T, H, W)
+        num_cond_latents: int = 0,    # Number of conditioning latent frames (for I2V)
+        return_kv: bool = False,      # Return K/V for caching
+        kv_cache: tuple | None = None,  # Pre-computed K/V cache
+        skip_crs_attn: bool = False,  # Skip cross-attention (for cache init)
         **kwargs
-    ) -> torch.Tensor:
+    ) -> torch.Tensor | tuple:
         """
         Forward pass with AdaLN modulation.
+        
+        Args:
+            num_cond_latents: For I2V, number of conditioning latent frames.
+                              These frames use split attention behavior.
+            return_kv: If True, return (x, (k_cache, v_cache))
+            kv_cache: Pre-computed K/V from conditioning frames
+            skip_crs_attn: If True, skip cross-attention (used during cache init)
         """
         B, N, C = x.shape
         T, H, W = latent_shape
@@ -592,17 +800,47 @@ class LongCatTransformerBlock(nn.Module):
         x_norm = modulate_fp32(self.norm_attn, x.view(B, T, -1, C), shift_msa, scale_msa)
         x_norm = x_norm.view(B, N, C)
 
-        attn_out = self.self_attn(x_norm, latent_shape=latent_shape)
+        # Handle KV cache
+        if kv_cache is not None:
+            # Move cache to device if offloaded
+            kv_cache = (kv_cache[0].to(x.device), kv_cache[1].to(x.device))
+            attn_out = self.self_attn.forward_with_kv_cache(
+                x_norm,
+                latent_shape=latent_shape,
+                num_cond_latents=num_cond_latents,
+                kv_cache=kv_cache,
+            )
+            kv_cache_new = None  # Don't return cache when using cache
+        else:
+            attn_result = self.self_attn(
+                x_norm, 
+                latent_shape=latent_shape,
+                num_cond_latents=num_cond_latents,
+                return_kv=return_kv,
+            )
+            if return_kv:
+                attn_out, kv_cache_new = attn_result
+            else:
+                attn_out = attn_result
+                kv_cache_new = None
 
         # Residual with gating (CRITICAL: FP32 like original, then cast back)
         with torch.amp.autocast(device_type='cuda', dtype=torch.float32):
             x = x + (gate_msa * attn_out.view(B, T, -1, C)).view(B, N, C)
         x = x.to(x_orig_dtype)
 
-        # === Cross-Attention ===
-        x_norm_cross = self.norm_cross(x)
-        cross_out = self.cross_attn(x_norm_cross, context)
-        x = x + cross_out
+        # === Cross-Attention (skip if requested) ===
+        if not skip_crs_attn:
+            x_norm_cross = self.norm_cross(x)
+            # When using KV cache, no need for num_cond_latents in cross-attn
+            cross_num_cond = 0 if kv_cache is not None else num_cond_latents
+            cross_out = self.cross_attn(
+                x_norm_cross, 
+                context,
+                latent_shape=latent_shape,
+                num_cond_latents=cross_num_cond
+            )
+            x = x + cross_out
 
         # === FFN ===
         x_norm_ffn = modulate_fp32(self.norm_ffn, x.view(B, T, -1, C), shift_mlp, scale_mlp)
@@ -615,6 +853,8 @@ class LongCatTransformerBlock(nn.Module):
             x = x + (gate_mlp * ffn_out.view(B, T, -1, C)).view(B, N, C)
         x = x.to(x_orig_dtype)
 
+        if return_kv:
+            return x, kv_cache_new
         return x
 
 
@@ -670,16 +910,12 @@ class FinalLayer(nn.Module):
         B, N, C = x.shape
         T, _, _ = latent_shape
 
-        # AdaLN modulation (FP32 for stability like original)
-        with torch.amp.autocast(device_type='cuda', dtype=torch.float32):
-            t_mod = self.adaln_act(t)
-            mod_params, _ = self.adaln_linear(t_mod)
-            # Ensure FP32 output (needed when LoRA is applied)
-            if mod_params.dtype != torch.float32:
-                mod_params = mod_params.float()
-            shift, scale = mod_params.unsqueeze(2).chunk(2, dim=-1)
+        # AdaLN modulation
+        t_mod = self.adaln_act(t)
+        mod_params, _ = self.adaln_linear(t_mod)
+        shift, scale = mod_params.unsqueeze(2).chunk(2, dim=-1)
 
-        # Modulate
+        # Modulate (converts to FP32 internally for stability)
         x = modulate_fp32(self.norm, x.view(B, T, -1, C), shift, scale)
         x = x.reshape(B, N, C)
 
@@ -696,8 +932,6 @@ class FinalLayer(nn.Module):
 class LongCatTransformer3DModel(CachableDiT):
     """
     Native LongCat Video Transformer using FastVideo layers.
-    
-    This is a Phase 2 implementation that replaces third_party dependencies.
     """
 
     # FSDP sharding: shard at each transformer block
@@ -789,13 +1023,28 @@ class LongCatTransformer3DModel(CachableDiT):
         encoder_attention_mask: torch.Tensor | None = None,  # [B, N_text]
         encoder_hidden_states_image: torch.Tensor | list[torch.Tensor] | None = None,
         guidance: float | None = None,         # Unused, for API compatibility
+        num_cond_latents: int = 0,             # For I2V: number of conditioning latent frames
+        # === KV Cache Parameters ===
+        return_kv: bool = False,               # If True, return (output, kv_cache_dict)
+        kv_cache_dict: dict | None = None,     # Pre-computed {block_idx: (k, v)}
+        skip_crs_attn: bool = False,           # Skip cross-attention (for cache init)
+        offload_kv_cache: bool = False,        # Move cache to CPU after compute
         **kwargs
-    ) -> torch.Tensor:
+    ) -> torch.Tensor | tuple[torch.Tensor, dict]:
         """
         Forward pass with FastVideo parameter ordering.
         
         NOTE: This follows FastVideo convention:
               (hidden_states, encoder_hidden_states, timestep)
+              
+        Args:
+            num_cond_latents: For I2V, number of conditioning latent frames.
+                              These frames are treated as "clean" (timestep=0)
+                              and use split attention behavior.
+            return_kv: If True, return (output, kv_cache_dict)
+            kv_cache_dict: Pre-computed K/V cache {block_idx: (k, v)}
+            skip_crs_attn: If True, skip cross-attention (for cache init)
+            offload_kv_cache: If True, move cache to CPU after compute
         """
         B, _, T, H, W = hidden_states.shape
 
@@ -825,12 +1074,31 @@ class LongCatTransformer3DModel(CachableDiT):
             encoder_attention_mask=encoder_attention_mask
         )  # [B, N_text, C]
 
-        # 4. Transformer blocks
+        # 4. Transformer blocks with optional KV cache
+        kv_cache_dict_ret = {} if return_kv else None
+        
         for i, block in enumerate(self.blocks):
-            x = block(
+            # Get cache for this block if available
+            block_kv_cache = kv_cache_dict.get(i, None) if kv_cache_dict else None
+            
+            block_out = block(
                 x, context, t,
-                latent_shape=(N_t, N_h, N_w)
+                latent_shape=(N_t, N_h, N_w),
+                num_cond_latents=num_cond_latents,
+                return_kv=return_kv,
+                kv_cache=block_kv_cache,
+                skip_crs_attn=skip_crs_attn,
             )
+            
+            if return_kv:
+                x, kv_cache = block_out
+                # Store cache
+                if offload_kv_cache:
+                    kv_cache_dict_ret[i] = (kv_cache[0].cpu(), kv_cache[1].cpu())
+                else:
+                    kv_cache_dict_ret[i] = (kv_cache[0].contiguous(), kv_cache[1].contiguous())
+            else:
+                x = block_out
 
         # 5. Output projection
         output = self.final_layer(x, t, latent_shape=(N_t, N_h, N_w))
@@ -841,6 +1109,8 @@ class LongCatTransformer3DModel(CachableDiT):
         # Cast to float32 for better accuracy (as per original)
         output = output.to(torch.float32)
 
+        if return_kv:
+            return output, kv_cache_dict_ret
         return output
 
     def unpatchify(self, x: torch.Tensor, N_t: int, N_h: int, N_w: int) -> torch.Tensor:

fastvideo/models/dits/wanvideo.py

@@ -734,18 +734,8 @@ class WanTransformer3DModel(CachableDiT):
                         block, hidden_states, encoder_hidden_states,
                         timestep_proj, freqs_cis, attention_mask)
             else:
-                offload_mgr = getattr(self, "_layerwise_offload_manager", None)
-                use_offload = offload_mgr is not None and getattr(offload_mgr, "enabled", False)
-
-                for i, block in enumerate(self.blocks):
-                    scope = offload_mgr.layer_scope(
-                        prefetch_layer_idx=i + 1 if i + 1 < len(self.blocks) else None,
-                        release_layer_idx=i,
-                        non_blocking=True,
-                    ) if use_offload else nullcontext()
-                    
-                    with scope:
-                        hidden_states = block(hidden_states, encoder_hidden_states,
+                for block in self.blocks:
+                    hidden_states = block(hidden_states, encoder_hidden_states,
                                               timestep_proj, freqs_cis, attention_mask)
             # if teacache is enabled, we need to cache the original hidden states
 

fastvideo/models/encoders/reason1.py

@@ -0,0 +1,353 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Reason1 (Qwen2.5-VL) text encoder."""
+
+import os
+from dataclasses import dataclass
+from collections.abc import Iterable
+
+import torch
+from transformers import AutoProcessor
+
+from fastvideo.configs.models.encoders import BaseEncoderOutput, Reason1Config
+from fastvideo.logger import init_logger
+from fastvideo.models.encoders.base import TextEncoder
+from fastvideo.models.loader.weight_utils import default_weight_loader
+from fastvideo.platforms import AttentionBackendEnum
+
+from fastvideo.models.encoders.qwen2_5_vl_custom import (
+    Qwen2_5_VLForConditionalGenerationSimple,
+    Qwen2_5_VLConfig,
+    get_rope_index,
+)
+
+logger = init_logger(__name__)
+
+
+@dataclass(frozen=True)
+class _WeightsSource:
+    """Mimic `TextEncoderLoader.Source` (avoid import cycles)."""
+
+    model_or_path: str
+    prefix: str = ""
+    fall_back_to_pt: bool = True
+    allow_patterns_overrides: list[str] | None = None
+
+
+
+
+class Reason1TextEncoder(TextEncoder):
+    """Reason1 (Qwen2.5-VL) text encoder."""
+
+    _supported_attention_backends: tuple[AttentionBackendEnum, ...] = (
+        AttentionBackendEnum.FLASH_ATTN,
+        AttentionBackendEnum.TORCH_SDPA,
+    )
+    
+    def __init__(self, config: Reason1Config, prefix: str = "", checkpoint_path: str | None = None):
+        super().__init__(config)
+        
+        self.prefix = prefix
+        self.quant_config = None  # For future quantization support
+        
+        self.embedding_concat_strategy = config.arch_config.embedding_concat_strategy
+        self.n_layers_per_group = config.arch_config.n_layers_per_group
+        self.num_embedding_padding_tokens = config.arch_config.num_embedding_padding_tokens
+        
+        config_path = checkpoint_path if checkpoint_path else config.tokenizer_type
+        
+        logger.info("Initializing Reason1TextEncoder (Qwen2.5-VL) from %s", config_path)
+        try:
+            from transformers import AutoConfig as HFAutoConfig
+            hf_config = HFAutoConfig.from_pretrained(
+                config_path,
+                trust_remote_code=True,
+            )
+        except Exception as e:
+            logger.warning("Failed to load HF config from %s (%s). Using default Qwen2.5-VL-7B config.",
+                           config_path, e)
+            hf_config = Qwen2_5_VLConfig(
+                    hidden_size=3584,
+                    intermediate_size=18944,
+                    max_window_layers=28,
+                    num_attention_heads=28,
+                    num_hidden_layers=28,
+                    num_key_value_heads=4,
+                    tie_word_embeddings=False,
+                    vocab_size=152064,
+            )
+        
+        hf_config.output_hidden_states = True
+        
+        if hasattr(config.arch_config, '_attn_implementation') and config.arch_config._attn_implementation:
+            hf_config._attn_implementation = config.arch_config._attn_implementation
+        else:
+            hf_config._attn_implementation = "flash_attention_2"
+        logger.info("Reason1 attention implementation: %s", getattr(hf_config, "_attn_implementation", None))
+        
+        with torch.device("meta"):
+            self.model = Qwen2_5_VLForConditionalGenerationSimple(hf_config)
+        
+        self.processor = AutoProcessor.from_pretrained(
+            config_path,
+            trust_remote_code=True,
+        )
+
+        weights_override = os.getenv("FASTVIDEO_REASON1_WEIGHTS_PATH")
+        if weights_override:
+            self.secondary_weights = (
+                _WeightsSource(
+                    model_or_path=weights_override,
+                    prefix="",
+                    fall_back_to_pt=True,
+                    allow_patterns_overrides=None,
+                ),
+            )
+            logger.info("Reason1TextEncoder: overlaying weights from %s", weights_override)
+        
+        self._weights_loaded = False
+    
+    def forward(
+        self,
+        input_ids: torch.Tensor | None,
+        position_ids: torch.Tensor | None = None,
+        attention_mask: torch.Tensor | None = None,
+        inputs_embeds: torch.Tensor | None = None,
+        output_hidden_states: bool | None = None,
+        **kwargs,
+    ) -> BaseEncoderOutput:
+        # Cosmos2.5 alignment: keep attention_mask=None.
+        outputs = self.model(
+            input_ids=input_ids,
+            attention_mask=None,
+            position_ids=position_ids,
+            inputs_embeds=inputs_embeds,
+            output_hidden_states=True,
+            return_dict=True,
+            pixel_values=kwargs.get('pixel_values', None),
+            pixel_values_videos=kwargs.get('pixel_values_videos', None),
+            image_grid_thw=kwargs.get('image_grid_thw', None),
+            video_grid_thw=kwargs.get('video_grid_thw', None),
+        )
+        
+        hidden_states = outputs.hidden_states
+        last_hidden_state = hidden_states[-1]
+        
+        return BaseEncoderOutput(
+            last_hidden_state=last_hidden_state,
+            hidden_states=hidden_states if output_hidden_states else None,
+            attention_mask=None,
+        )
+    
+    def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
+        first_weight = None
+        weights_list = []
+        for name, weight in weights:
+            if first_weight is None:
+                first_weight = weight
+                self.model = self.model.to_empty(device=weight.device)
+                self.model.init_weights(buffer_device=weight.device)
+            weights_list.append((name, weight))
+        
+        params_dict = dict(self.model.named_parameters())
+        loaded_params: set[str] = set()
+        skipped_weights = {"lm_head": 0, "visual": 0, "decoder": 0}
+        
+        for name, loaded_weight in weights_list:
+            if "lm_head" in name:
+                skipped_weights["lm_head"] += 1
+                continue
+            if "visual" in name:
+                skipped_weights["visual"] += 1
+                continue
+            if "decoder" in name:
+                skipped_weights["decoder"] += 1
+                continue
+            
+            # Handle stacked params mapping (for quantized models)
+            for param_name, weight_name, shard_id in self.config.arch_config.stacked_params_mapping:
+                if weight_name not in name:
+                    continue
+                name = name.replace(weight_name, param_name)
+                if name.endswith(".bias") and name not in params_dict:
+                    continue
+                
+                if name not in params_dict:
+                    continue
+                
+                param = params_dict[name]
+                weight_loader = param.weight_loader
+                weight_loader(param, loaded_weight, shard_id)
+                param_name_with_prefix = f"model.{name}" if self.prefix == "" else f"{self.prefix}.{name}"
+                loaded_params.add(param_name_with_prefix)
+                break
+            else:
+                if name.endswith(".bias") and name not in params_dict:
+                    continue
+                
+                if name not in params_dict:
+                    continue
+                
+                param = params_dict[name]
+                weight_loader = getattr(param, "weight_loader", default_weight_loader)
+                weight_loader(param, loaded_weight)
+                param_name_with_prefix = f"model.{name}" if self.prefix == "" else f"{self.prefix}.{name}"
+                loaded_params.add(param_name_with_prefix)
+        if first_weight is not None:
+            self.model = self.model.to(first_weight.device)
+        
+        all_params = set(f"model.{name}" if self.prefix == "" else f"{self.prefix}.{name}" 
+                        for name in params_dict.keys())
+        loaded_params.update(all_params)
+        
+        # Mark weights as loaded
+        self._weights_loaded = True
+        return loaded_params
+    
+    def compute_text_embeddings_online(
+        self,
+        data_batch: dict[str, list[str]],
+        input_caption_key: str,
+    ) -> torch.Tensor:
+        prompts = data_batch[input_caption_key]
+        return self.compute_text_embeddings(prompts)
+    
+    def compute_text_embeddings(
+        self,
+        prompts: list[str],
+        device: str | torch.device = "cuda",
+    ) -> torch.Tensor:
+        """Compute embeddings for a list of prompts."""
+        input_ids_batch = []
+        
+        tok = getattr(self.processor, "tokenizer", None)
+        if tok is None:
+            raise RuntimeError("Reason1TextEncoder requires processor.tokenizer")
+        pad_id = getattr(tok, "pad_id", None)
+        if pad_id is None:
+            pad_id = getattr(tok, "pad_token_id", None)
+        if pad_id is None:
+            pad_id = getattr(self.model.config, "pad_token_id", None)
+        if pad_id is None:
+            pad_id = 0
+        
+        for prompt in prompts:
+            conversations = [
+                {
+                    "role": "system",
+                    "content": [
+                        {
+                            "type": "text",
+                            "text": "You are a helpful assistant who will provide prompts to an image generator.",
+                        }
+                    ],
+                },
+                {
+                    "role": "user",
+                    "content": [
+                        {
+                            "type": "text",
+                            "text": prompt,
+                        }
+                    ],
+                },
+            ]
+            
+            try:
+                tokenizer_output = tok.apply_chat_template(
+                    conversations,
+                    tokenize=True,
+                    add_generation_prompt=False,
+                    add_vision_id=False,
+                )
+            except TypeError:
+                tokenizer_output = tok.apply_chat_template(
+                conversations,
+                tokenize=True,
+                add_generation_prompt=False,
+            )
+
+            if isinstance(tokenizer_output, dict) and "input_ids" in tokenizer_output:
+                input_ids = tokenizer_output["input_ids"]
+                if hasattr(input_ids, "tolist"):
+                    input_ids = input_ids.tolist()
+            else:
+                input_ids = tokenizer_output
+                if hasattr(input_ids, "tolist"):
+                    input_ids = input_ids.tolist()
+                if isinstance(input_ids, list) and len(input_ids) == 1 and isinstance(
+                        input_ids[0], list):
+                    input_ids = input_ids[0]
+                if not isinstance(input_ids, list):
+                    raise RuntimeError(
+                        f"Unexpected chat_template output type: {type(tokenizer_output)}"
+                    )
+            
+            if self.num_embedding_padding_tokens > len(input_ids):
+                pad_len = self.num_embedding_padding_tokens - len(input_ids)
+                input_ids = input_ids + [pad_id] * pad_len
+            else:
+                input_ids = input_ids[:self.num_embedding_padding_tokens]
+            
+            input_ids = torch.LongTensor(input_ids).to(device=device)
+            input_ids_batch.append(input_ids)
+        
+        input_ids_batch = torch.stack(input_ids_batch, dim=0)
+
+        # Cosmos2.5 alignment: keep attention_mask=None.
+        target_device = input_ids_batch.device
+        try:
+            embed_device = self.model.model.embed_tokens.weight.device  # type: ignore[attr-defined]
+        except Exception:
+            embed_device = None
+        if embed_device is not None and embed_device != target_device:
+            self.model = self.model.to(target_device)
+        
+        with torch.no_grad():
+            position_ids, _ = get_rope_index(
+                self.model.config,
+                input_ids_batch,
+                image_grid_thw=None,
+                video_grid_thw=None,
+                second_per_grid_ts=None,
+                attention_mask=None,
+            )
+            position_ids = position_ids.to(target_device)
+                
+            outputs = self.model.model(
+                input_ids=input_ids_batch,
+                position_ids=position_ids,
+                attention_mask=None,
+                output_hidden_states=True,
+                return_dict=True,
+                use_cache=False,
+            )
+            hidden_states = outputs.hidden_states
+        
+        normalized_hidden_states = []
+        for layer_idx in range(1, len(hidden_states)):
+            normalized_state = self._mean_normalize(hidden_states[layer_idx])
+            normalized_hidden_states.append(normalized_state)
+        
+        if self.embedding_concat_strategy == "full_concat":
+            text_embeddings = torch.cat(normalized_hidden_states, dim=-1)
+        elif self.embedding_concat_strategy == "mean_pooling":
+            text_embeddings = torch.stack(normalized_hidden_states).mean(dim=0)
+        elif self.embedding_concat_strategy == "pool_every_n_layers_and_concat":
+            pooled_embeddings = []
+            for i in range(0, len(normalized_hidden_states), self.n_layers_per_group):
+                group = normalized_hidden_states[i : i + self.n_layers_per_group]
+                pooled = torch.stack(group).mean(dim=0)
+                pooled_embeddings.append(pooled)
+            text_embeddings = torch.cat(pooled_embeddings, dim=-1)
+        else:
+            raise ValueError(
+                f"Unknown embedding_concat_strategy: {self.embedding_concat_strategy}"
+            )
+        
+        return text_embeddings
+    @staticmethod
+    def _mean_normalize(tensor: torch.Tensor) -> torch.Tensor:
+        return (tensor - tensor.mean(dim=-1, keepdim=True)) / (
+            tensor.std(dim=-1, keepdim=True) + 1e-8
+        )
+

fastvideo/models/loader/component_loader.py

@@ -15,8 +15,7 @@ import torch.distributed as dist
 import torch.nn as nn
 from safetensors.torch import load_file as safetensors_load_file
 from torch.distributed import init_device_mesh
-from transformers import AutoImageProcessor, AutoTokenizer
-from transformers import UMT5EncoderModel
+from transformers import AutoImageProcessor, AutoProcessor, AutoTokenizer
 from transformers.utils import SAFE_WEIGHTS_INDEX_NAME
 
 from fastvideo.configs.models import EncoderConfig
@@ -35,8 +34,8 @@ from fastvideo.models.loader.weight_utils import (
     safetensors_weights_iterator,
 )
 from fastvideo.models.registry import ModelRegistry
-from fastvideo.utils import PRECISION_TO_TYPE
-from fastvideo.models.layerwise_offload import LayerwiseOffloadManager
+from fastvideo.utils import PRECISION_TO_TYPE, is_pin_memory_available
+from fastvideo.hooks.layerwise_offload import enable_layerwise_offload
 
 logger = init_logger(__name__)
 
@@ -91,10 +90,12 @@ class ComponentLoader(ABC):
 
         if module_type in module_loaders:
             loader_cls, expected_library = module_loaders[module_type]
-            # Assert that the library matches what's expected for this module type
-            assert transformers_or_diffusers == expected_library, (
-                f"{module_type} must be loaded from {expected_library}, got {transformers_or_diffusers}"
-            )
+            # Allow fastvideo.* libraries for custom implementations (e.g. Cosmos2_5Pipeline)
+            # that aren't available in diffusers/transformers yet
+            is_fastvideo_module = transformers_or_diffusers.startswith("fastvideo.")
+            if not is_fastvideo_module:
+                # Assert that the library matches what's expected for this module type
+                assert transformers_or_diffusers == expected_library, f"{module_type} must be loaded from {expected_library}, got {transformers_or_diffusers}"
             return loader_cls()
 
         # For unknown module types, use a generic loader
@@ -279,7 +280,7 @@ class TextEncoderLoader(ComponentLoader):
         target_device: torch.device,
         fastvideo_args: FastVideoArgs,
         dtype: str = "fp16",
-        use_text_encoder_override: bool = False, # prevent subclasses from misusing
+        use_text_encoder_override: bool = False,  # prevent subclasses from misusing
     ):
         use_cpu_offload = (
             fastvideo_args.text_encoder_cpu_offload
@@ -296,7 +297,10 @@ class TextEncoderLoader(ComponentLoader):
             )
 
         # Set quantization config if specified
-        if use_text_encoder_override and fastvideo_args.override_text_encoder_quant is not None:
+        if (
+            use_text_encoder_override
+            and fastvideo_args.override_text_encoder_quant is not None
+        ):
             if fastvideo_args.override_text_encoder_safetensors is None:
                 raise ValueError(
                     "override_text_encoder_quant is set but override_text_encoder_safetensors is None"
@@ -313,7 +317,10 @@ class TextEncoderLoader(ComponentLoader):
                 model: TextEncoder = model_cls(model_config)  # type: ignore
 
             weights_to_load = {name for name, _ in model.named_parameters()}
-            if use_text_encoder_override and fastvideo_args.override_text_encoder_safetensors is not None:
+            if (
+                use_text_encoder_override
+                and fastvideo_args.override_text_encoder_safetensors is not None
+            ):
                 loaded_weights: set[str] = model.load_weights(
                     safetensors_weights_iterator(
                         [fastvideo_args.override_text_encoder_safetensors],
@@ -340,6 +347,7 @@ class TextEncoderLoader(ComponentLoader):
             from fastvideo.platforms import current_platform
 
             if use_cpu_offload:
+                pin_cpu_memory = fastvideo_args.pin_cpu_memory and is_pin_memory_available()
                 # Disable FSDP for MPS as it's not compatible
                 if current_platform.is_mps():
                     logger.info(
@@ -357,7 +365,7 @@ class TextEncoderLoader(ComponentLoader):
                         reshard_after_forward=True,
                         mesh=mesh["offload"],
                         fsdp_shard_conditions=model._fsdp_shard_conditions,
-                        pin_cpu_memory=fastvideo_args.pin_cpu_memory,
+                        pin_cpu_memory=pin_cpu_memory,
                     )
                 else:
                     mesh = init_device_mesh(
@@ -371,7 +379,7 @@ class TextEncoderLoader(ComponentLoader):
                         reshard_after_forward=True,
                         mesh=mesh["offload"],
                         fsdp_shard_conditions=model._fsdp_shard_conditions,
-                        pin_cpu_memory=fastvideo_args.pin_cpu_memory,
+                        pin_cpu_memory=pin_cpu_memory,
                     )
             # We only enable strict check for non-quantized models
             # that have loaded weights tracking currently.
@@ -449,6 +457,33 @@ class TokenizerLoader(ComponentLoader):
         """Load the tokenizer based on the model path, and inference args."""
         logger.info("Loading tokenizer from %s", model_path)
 
+        # Cosmos2.5 stores an AutoProcessor config in `tokenizer/config.json` (not a tokenizer
+        # config). Use its `_name_or_path` (e.g. Qwen/Qwen2.5-VL-7B-Instruct) as the source.
+        tokenizer_cfg_path = os.path.join(model_path, "config.json")
+        if os.path.exists(tokenizer_cfg_path):
+            try:
+                with open(tokenizer_cfg_path, "r") as f:
+                    tokenizer_cfg = json.load(f)
+                if isinstance(tokenizer_cfg, dict) and (
+                    tokenizer_cfg.get("_class_name") == "AutoProcessor"
+                    or "processor_type" in tokenizer_cfg
+                ):
+                    src = tokenizer_cfg.get("_name_or_path", "")
+                    if isinstance(src, str) and src.strip():
+                        processor = AutoProcessor.from_pretrained(
+                            src.strip(),
+                            trust_remote_code=True,
+                        )
+                        logger.info(
+                            "Loaded tokenizer/processor from %s: %s",
+                            src,
+                            processor.__class__.__name__,
+                        )
+                        return processor
+            except Exception:
+                # If parsing fails, fall through to AutoTokenizer below.
+                pass
+
         tokenizer = AutoTokenizer.from_pretrained(
             model_path,  # "<path to model>/tokenizer"
             # in v0, this was same string as encoder_name "ClipTextModel"
@@ -491,19 +526,40 @@ class VAELoader(ComponentLoader):
             if fastvideo_args.pipeline_config.vae_precision
             else torch.bfloat16
         ):
+            # Cosmos2.5 uses a Wan2.1 VAE stored as `tokenizer.safetensors` under the VAE folder.
+            is_cosmos25 = fastvideo_args.pipeline_config.__class__.__name__ == "Cosmos25Config"
+            if class_name == "AutoencoderKLWan" and is_cosmos25:
+                from fastvideo.models.vaes.cosmos25wanvae import Cosmos25WanVAE
+
+                dtype = PRECISION_TO_TYPE[fastvideo_args.pipeline_config.vae_precision]
+                vae = Cosmos25WanVAE(device=target_device, dtype=dtype)
+
+                weight_path = os.path.join(model_path, "tokenizer.safetensors")
+                if not os.path.exists(weight_path):
+                    raise FileNotFoundError(
+                        f"Missing Cosmos2.5 VAE weights: {weight_path}"
+                    )
+                sd = safetensors_load_file(weight_path)
+                vae.load_state_dict(sd, strict=False)
+                return vae.eval()
+
             vae_cls, _ = ModelRegistry.resolve_model_cls(class_name)
             vae = vae_cls(vae_config).to(target_device)
 
         # Find all safetensors files
         safetensors_list = glob.glob(
-            os.path.join(str(model_path), "*.safetensors")
-        )
-        loaded = {}
-        for sf_file in safetensors_list:
-            loaded.update(safetensors_load_file(sf_file))
-        vae.load_state_dict(
-            loaded, strict=False
-        )  # We might only load encoder or decoder
+            os.path.join(str(model_path), "*.safetensors"))
+        if not safetensors_list:
+            raise ValueError(f"No safetensors files found in {model_path}")
+        # Common case: a single `.safetensors` checkpoint file.
+        # Some models may be sharded into multiple files; in that case we merge.
+        if len(safetensors_list) == 1:
+            loaded = safetensors_load_file(safetensors_list[0])
+        else:
+            loaded = {}
+            for sf_file in safetensors_list:
+                loaded.update(safetensors_load_file(sf_file))
+        vae.load_state_dict(loaded, strict=False)
 
         return vae.eval()
 
@@ -581,7 +637,17 @@ class TransformerLoader(ComponentLoader):
         ]
 
         # Load the model using FSDP loader
+        logger.info("Loading model from %s, default_dtype: %s", cls_name,
+                    default_dtype)
         assert fastvideo_args.hsdp_shard_dim is not None
+        # Cosmos2.5 checkpoints can include extra entries not present in the
+        # instantiated model (e.g. pos_embedder ranges / *_extra_state). Load
+        # non-strictly for Cosmos2.5 only; keep upstream strict behavior for others.
+        strict_load = not (
+            cls_name.startswith("Cosmos25")
+            or cls_name == "Cosmos25Transformer3DModel"
+            or getattr(fastvideo_args.pipeline_config, "prefix", "") == "Cosmos25"
+        )
         model = maybe_load_fsdp_model(
             model_cls=model_cls,
             init_params={"config": dit_config, "hf_config": hf_config},
@@ -589,6 +655,7 @@ class TransformerLoader(ComponentLoader):
             device=get_local_torch_device(),
             hsdp_replicate_dim=fastvideo_args.hsdp_replicate_dim,
             hsdp_shard_dim=fastvideo_args.hsdp_shard_dim,
+            strict=strict_load,
             cpu_offload=fastvideo_args.dit_cpu_offload,
             pin_cpu_memory=fastvideo_args.pin_cpu_memory,
             fsdp_inference=fastvideo_args.use_fsdp_inference,
@@ -611,35 +678,8 @@ class TransformerLoader(ComponentLoader):
 
         model = model.eval()
 
-        if fastvideo_args.dit_layerwise_offload and hasattr(model, "blocks"):
-            # Check if this is a Wan model (only Wan models support layerwise offload)
-            is_wan_model = "Wan" in cls_name
-            if not is_wan_model:
-                logger.warning(
-                    "Layerwise offload is currently only supported for Wan models. "
-                    "Model class '%s' does not support layerwise offload. "
-                    "Disabling layerwise offload for this model.",
-                    cls_name
-                )
-            else:
-                try:
-                    num_layers = len(getattr(model, "blocks"))
-                except TypeError:
-                    num_layers = None
-                if isinstance(num_layers, int) and num_layers > 0:
-                    # Ensure model is on the correct device (CUDA) before initializing manager
-                    # This ensures non-managed parameters (embeddings, final norms) are on GPU
-                    model = model.to(get_local_torch_device())
-                    mgr = LayerwiseOffloadManager(
-                        model,
-                        module_list_attr="blocks",
-                        num_layers=num_layers,
-                        enabled=True,
-                        pin_cpu_memory=fastvideo_args.pin_cpu_memory,
-                        auto_initialize=True,
-                    )
-                    setattr(model, "_layerwise_offload_manager", mgr)
-
+        if fastvideo_args.inference_mode and fastvideo_args.dit_layerwise_offload:
+            enable_layerwise_offload(model)
         return model
 
 

fastvideo/models/loader/fsdp_load.py

@@ -23,7 +23,7 @@ from fastvideo.logger import init_logger
 from fastvideo.models.loader.utils import (get_param_names_mapping,
                                            hf_to_custom_state_dict)
 from fastvideo.models.loader.weight_utils import safetensors_weights_iterator
-from fastvideo.utils import set_mixed_precision_policy
+from fastvideo.utils import set_mixed_precision_policy, is_pin_memory_available
 
 logger = init_logger(__name__)
 
@@ -67,6 +67,7 @@ def maybe_load_fsdp_model(
     default_dtype: torch.dtype,
     param_dtype: torch.dtype,
     reduce_dtype: torch.dtype,
+    strict: bool = True,
     cpu_offload: bool = False,
     fsdp_inference: bool = False,
     output_dtype: torch.dtype | None = None,
@@ -106,6 +107,7 @@ def maybe_load_fsdp_model(
         logger.info("Disabling FSDP for MPS platform as it's not compatible")
 
     if use_fsdp:
+        pin_cpu_memory = pin_cpu_memory and is_pin_memory_available()
         world_size = hsdp_replicate_dim * hsdp_shard_dim
         if not training_mode and not fsdp_inference:
             hsdp_replicate_dim = world_size
@@ -141,7 +143,7 @@ def maybe_load_fsdp_model(
         weight_iterator,
         device,
         default_dtype,
-        strict=True,
+        strict=strict,
         cpu_offload=cpu_offload,
         param_names_mapping=param_names_mapping_fn,
     )
@@ -151,6 +153,7 @@ def maybe_load_fsdp_model(
                 f"Unexpected param or buffer {n} on meta device.")
         # Avoid unintended computation graph accumulation during inference
         if isinstance(p, torch.nn.Parameter):
+            
             p.requires_grad = False
 
     compile_in_loader = enable_torch_compile and training_mode
@@ -293,6 +296,26 @@ def load_model_from_full_model_state_dict(
     for target_param_name, full_tensor in custom_param_sd.items():
         meta_sharded_param = meta_sd.get(target_param_name)
         if meta_sharded_param is None:
+            # Some checkpoints include extra entries that are not part of the
+            # instantiated model's state_dict (e.g. `_extra_state` keys from
+            # some FSDP checkpoint formats). These can be safely skipped.
+            if (target_param_name.endswith("._extra_state")
+                    or target_param_name.endswith("_extra_state")):
+                logger.warning(
+                    "Skipping non-parameter checkpoint key: %s",
+                    target_param_name,
+                )
+                continue
+
+            # For non-strict loads, treat this as an "unexpected key" and skip it
+            # (mirrors torch.nn.Module.load_state_dict(strict=False)).
+            if not strict:
+                logger.warning(
+                    "Skipping unexpected checkpoint key (not present in model): %s",
+                    target_param_name,
+                )
+                continue
+
             raise ValueError(
                 f"Parameter {target_param_name} not found in custom model state dict. The hf to custom mapping may be incorrect."
             )

fastvideo/models/registry.py

@@ -30,6 +30,7 @@ _TEXT_TO_VIDEO_DIT_MODELS = {
     "CausalWanTransformer3DModel": ("dits", "causal_wanvideo", "CausalWanTransformer3DModel"),
     "StepVideoModel": ("dits", "stepvideo", "StepVideoModel"),
     "CosmosTransformer3DModel": ("dits", "cosmos", "CosmosTransformer3DModel"),
+    "Cosmos25Transformer3DModel": ("dits", "cosmos2_5", "Cosmos25Transformer3DModel"),
     "LongCatVideoTransformer3DModel": ("dits", "longcat_video_dit", "LongCatVideoTransformer3DModel"),  # Wrapper (Phase 1)
     "LongCatTransformer3DModel": ("dits", "longcat", "LongCatTransformer3DModel"),  # Native (Phase 2)
 }
@@ -50,6 +51,9 @@ _TEXT_ENCODER_MODELS = {
     "STEP1TextEncoder": ("encoders", "stepllm", "STEP1TextEncoder"),
     "BertModel": ("encoders", "clip", "CLIPTextModel"),
     "Qwen2_5_VLTextModel": ("encoders", "qwen2_5", "Qwen2_5_VLTextModel"),
+    "Reason1TextEncoder": ("encoders", "reason1", "Reason1TextEncoder"),
+    "Qwen2_5_VLForConditionalGeneration":
+    ("encoders", "reason1", "Reason1TextEncoder"),
 }
 
 _IMAGE_ENCODER_MODELS: dict[str, tuple] = {
@@ -72,6 +76,8 @@ _SCHEDULERS = {
      "FlowMatchEulerDiscreteScheduler"),
     "UniPCMultistepScheduler":
     ("schedulers", "scheduling_unipc_multistep", "UniPCMultistepScheduler"),
+    "FlowUniPCMultistepScheduler":
+    ("schedulers", "scheduling_flow_unipc_multistep", "FlowUniPCMultistepScheduler"),
     "SelfForcingFlowMatchScheduler":
     ("schedulers", "scheduling_self_forcing_flow_match",
      "SelfForcingFlowMatchScheduler"),

fastvideo/models/schedulers/scheduling_flow_unipc_multistep.py

@@ -109,6 +109,9 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
         sigmas = 1.0 - alphas
         sigmas = torch.from_numpy(sigmas).to(dtype=torch.float32)
 
+        # Needed when final_sigmas_type == "sigma_min" (kept for compatibility).
+        self.alphas_cumprod = torch.from_numpy(alphas).to(dtype=torch.float32)
+
         if not use_dynamic_shifting:
             # when use_dynamic_shifting is True, we apply the timestep shifting on the fly based on the image resolution
             assert shift is not None
@@ -171,6 +174,8 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
         sigmas: list[float] | None = None,
         mu: float | None | None = None,
         shift: float | None | None = None,
+        use_karras_sigmas: bool | None = None,
+        use_kerras_sigma: bool | None = None,
     ):
         """
         Sets the discrete timesteps used for the diffusion chain (to be run before inference).
@@ -186,21 +191,44 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
                 " you have to pass a value for `mu` when `use_dynamic_shifting` is set to be `True`"
             )
 
-        if sigmas is None:
-            assert num_inference_steps is not None
-            sigmas = np.linspace(self.sigma_max, self.sigma_min,
-                                 num_inference_steps +
-                                 1).copy()[:-1]  # pyright: ignore
+
+        # Cosmos official uses `use_kerras_sigma=True` and a specific EDM sigma schedule.
+        # Some external code uses the misspelling `use_kerras_sigma`; support both.
+        if use_karras_sigmas is None and use_kerras_sigma is not None:
+            use_karras_sigmas = use_kerras_sigma
+
+        if use_karras_sigmas:
+            # Force to use the exact sigma used in official EDM sampler:
+            # sigma_max=200, sigma_min=0.01, rho=7
+            sigma_max = 200.0
+            sigma_min = 0.01
+            rho = 7.0
+            # Match the official Cosmos implementation: Karras/EDM schedule with
+            # `num_inference_steps + 1` points, then `final_sigmas_type="zero"`
+            # appends the terminal sigma (0.0).
+            ramp = np.arange(num_inference_steps + 1,
+                             dtype=np.float32) / float(num_inference_steps)
+            min_inv_rho = sigma_min**(1 / rho)
+            max_inv_rho = sigma_max**(1 / rho)
+            sigmas = (max_inv_rho + ramp * (min_inv_rho - max_inv_rho))**rho
+            # Convert EDM sigma to flow-matching sigma in [0, 1).
+            sigmas = sigmas / (1.0 + sigmas)
+        else:
+            if sigmas is None:
+                assert num_inference_steps is not None
+                sigmas = np.linspace(self.sigma_max, self.sigma_min,
+                                     num_inference_steps +
+                                     1).copy()[:-1]  # pyright: ignore
 
         if self.config.use_dynamic_shifting:
             assert mu is not None
             sigmas = self.time_shift(mu, 1.0, sigmas)  # pyright: ignore
         else:
-            if shift is None:
-                shift = self.config.shift
-            assert isinstance(sigmas, np.ndarray)
-            sigmas = shift * sigmas / (1 +
-                                       (shift - 1) * sigmas)  # pyright: ignore
+            if not use_karras_sigmas:
+                if shift is None:
+                    shift = self.config.shift
+                assert isinstance(sigmas, np.ndarray)
+                sigmas = shift * sigmas / (1 + (shift - 1) * sigmas)  # pyright: ignore
 
         if self.config.final_sigmas_type == "sigma_min":
             sigma_last = ((1 - self.alphas_cumprod[0]) /
@@ -418,8 +446,12 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
         alpha_t, sigma_t = self._sigma_to_alpha_sigma_t(sigma_t)
         alpha_s0, sigma_s0 = self._sigma_to_alpha_sigma_t(sigma_s0)
 
-        lambda_t = torch.log(alpha_t) - torch.log(sigma_t)
-        lambda_s0 = torch.log(alpha_s0) - torch.log(sigma_s0)
+        # Numerical safety
+        eps = 1e-12
+        lambda_t = torch.log(torch.clamp(alpha_t, min=eps)) - torch.log(
+            torch.clamp(sigma_t, min=eps))
+        lambda_s0 = torch.log(torch.clamp(alpha_s0, min=eps)) - torch.log(
+            torch.clamp(sigma_s0, min=eps))
 
         h = lambda_t - lambda_s0
         device = sample.device
@@ -430,7 +462,8 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
             si = self.step_index - i  # pyright: ignore
             mi = model_output_list[-(i + 1)]
             alpha_si, sigma_si = self._sigma_to_alpha_sigma_t(self.sigmas[si])
-            lambda_si = torch.log(alpha_si) - torch.log(sigma_si)
+            lambda_si = torch.log(torch.clamp(alpha_si, min=eps)) - torch.log(
+                torch.clamp(sigma_si, min=eps))
             rk = (lambda_si - lambda_s0) / h
             rks.append(rk)
             assert mi is not None
@@ -563,8 +596,11 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
         alpha_t, sigma_t = self._sigma_to_alpha_sigma_t(sigma_t)
         alpha_s0, sigma_s0 = self._sigma_to_alpha_sigma_t(sigma_s0)
 
-        lambda_t = torch.log(alpha_t) - torch.log(sigma_t)
-        lambda_s0 = torch.log(alpha_s0) - torch.log(sigma_s0)
+        eps = 1e-12
+        lambda_t = torch.log(torch.clamp(alpha_t, min=eps)) - torch.log(
+            torch.clamp(sigma_t, min=eps))
+        lambda_s0 = torch.log(torch.clamp(alpha_s0, min=eps)) - torch.log(
+            torch.clamp(sigma_s0, min=eps))
 
         h = lambda_t - lambda_s0
         device = this_sample.device
@@ -575,7 +611,8 @@ class FlowUniPCMultistepScheduler(SchedulerMixin, ConfigMixin, BaseScheduler):
             si = self.step_index - (i + 1)  # pyright: ignore
             mi = model_output_list[-(i + 1)]
             alpha_si, sigma_si = self._sigma_to_alpha_sigma_t(self.sigmas[si])
-            lambda_si = torch.log(alpha_si) - torch.log(sigma_si)
+            lambda_si = torch.log(torch.clamp(alpha_si, min=eps)) - torch.log(
+                torch.clamp(sigma_si, min=eps))
             rk = (lambda_si - lambda_s0) / h
             rks.append(rk)
             assert mi is not None

fastvideo/models/vaes/cosmos25wanvae.py

@@ -0,0 +1,735 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: Apache-2.0
+"""
+Cosmos 2.5 / Wan2.1 VAE adapter.
+
+Why this exists:
+- Cosmos2.5 uses a Wan2.1-style VAE, but the *diffusion model* operates in a
+  **normalized latent space**:
+    z_norm = (z - mean) / std
+
+  Meanwhile, FastVideo's `AutoencoderKLWan` operates in the VAE's native latent
+  space (denormalized):
+    z = z_norm * std + mean
+
+This adapter provides a single, stable interface for FastVideo pipelines:
+- `encode(x)` returns an object with `.mean` / `.sample()` / `.mode()`
+- `decode(z)` returns a tensor in pixel space
+
+It also exposes flags used by pipeline stages to avoid double (de)normalization:
+- `handles_latent_norm = True`   -> stages should NOT normalize encoder latents
+- `handles_latent_denorm = True` -> stages should NOT denormalize before decode
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+from typing import Any, Optional
+
+import torch
+import torch.nn as nn
+import torch.nn.functional as F
+from einops import rearrange
+
+
+@dataclass
+class _TensorLatentDist:
+    """Minimal distribution-like wrapper used by pipeline stages."""
+
+    mean: torch.Tensor
+
+    def mode(self) -> torch.Tensor:
+        return self.mean
+
+    def sample(self, generator: Any | None = None) -> torch.Tensor:  # generator for API compatibility
+        # The official interface encodes deterministically; for compatibility we
+        # return the mean. (Stochastic posterior sampling isn't required for
+        # Cosmos2.5 inference.)
+        _ = generator
+        return self.mean
+
+
+class Cosmos25WanVAEAdapter(nn.Module):
+    """
+    Adapter that makes a Wan2.1-style VAE follow Cosmos2.5's latent contract:
+    - `encode()` returns **normalized** latents
+    - `decode()` expects **normalized** latents
+    """
+
+    # Pipeline stage hints (see latent_preparation.py / decoding.py / image_encoding.py)
+    handles_latent_norm: bool = True
+    handles_latent_denorm: bool = True
+    latent_norm_mode: str = "internal"  # informational
+
+    def __init__(
+        self,
+        inner: Any,
+        *,
+        latents_mean: Optional[torch.Tensor] = None,
+        latents_std: Optional[torch.Tensor] = None,
+    ) -> None:
+        super().__init__()
+        self.inner = inner
+
+        # Preserve `config` when available; some pipeline utilities expect it.
+        self.config = getattr(inner, "config", None)
+
+        # If not provided, try to derive from `config.latents_mean/std`.
+        cfg = self.config
+        if latents_mean is None and cfg is not None and hasattr(cfg, "latents_mean"):
+            latents_mean = torch.tensor(cfg.latents_mean, dtype=torch.float32).view(1, -1, 1, 1, 1)
+        if latents_std is None and cfg is not None and hasattr(cfg, "latents_std"):
+            latents_std = torch.tensor(cfg.latents_std, dtype=torch.float32).view(1, -1, 1, 1, 1)
+
+        if latents_mean is None or latents_std is None:
+            raise RuntimeError(
+                "Cosmos25WanVAEAdapter requires latents_mean/latents_std (either passed explicitly or available on inner.config)."
+            )
+
+        # Register as buffers so `.to(...)` moves them with the module.
+        self.register_buffer("_latents_mean", latents_mean, persistent=False)
+        self.register_buffer("_latents_std", latents_std, persistent=False)
+
+    def _to_latent_stats(self, like: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+        mean = self._latents_mean.to(device=like.device, dtype=like.dtype)
+        std = self._latents_std.to(device=like.device, dtype=like.dtype)
+        return mean, std
+
+    def get_latent_num_frames(self, num_pixel_frames: int) -> int:
+        # Keep parity with official interface.
+        if hasattr(self.inner, "get_latent_num_frames"):
+            return int(self.inner.get_latent_num_frames(num_pixel_frames))
+        return 1 + (num_pixel_frames - 1) // 4
+
+    def encode(self, x: torch.Tensor) -> _TensorLatentDist:
+        """
+        Returns *normalized* latents (Cosmos contract).
+        """
+        enc_out = self.inner.encode(x)
+
+        # Support common encoder output shapes:
+        # - DiagonalGaussianDistribution (FastVideo VAE): has `.mean` / `.sample()` / `.mode()`
+        # - diffusers EncoderOutput: has `.latent_dist`
+        # - raw tensor
+        if hasattr(enc_out, "latent_dist"):
+            dist = enc_out.latent_dist
+            z_mean = dist.mode() if hasattr(dist, "mode") else dist.mean
+        elif hasattr(enc_out, "mode") and hasattr(enc_out, "mean"):
+            z_mean = enc_out.mode()
+        elif isinstance(enc_out, torch.Tensor):
+            z_mean = enc_out
+        else:
+            attrs = [a for a in dir(enc_out) if not a.startswith("_")]
+            raise RuntimeError(
+                f"Unsupported VAE encoder output type: {type(enc_out)}. attrs={attrs}"
+            )
+
+        mean, std = self._to_latent_stats(z_mean)
+        z_norm = (z_mean - mean) / std
+        return _TensorLatentDist(z_norm)
+
+    def decode(self, z: torch.Tensor) -> torch.Tensor:
+        """
+        Expects *normalized* latents (Cosmos contract).
+        """
+        mean, std = self._to_latent_stats(z)
+        z_denorm = z * std + mean
+        out = self.inner.decode(z_denorm)
+        return out.sample if hasattr(out, "sample") else out
+
+
+#
+# Official-like Wan2.1 VAE implementation (ported from cosmos_predict2 wan2pt1.py)
+# -------------------------------------------------------------------------------
+# Motivation:
+# - We already solved checkpoint *key mapping* and can load official weights.
+# - Remaining output drift vs the official tokenizer is largely decoder-side.
+# - FastVideo's `AutoencoderKLWan` uses a different temporal upsample path
+#   (`DupUp3D` + `first_chunk` slicing), while the official tokenizer uses
+#   `Resample(mode="upsample3d")` with a time-conv + interleave reshape.
+#
+# This section ports the core modules (CausalConv3d/Resample/etc.) so we can run
+# a VAE that is behaviorally closer to the official implementation WITHOUT
+# importing any official repo classes at runtime.
+#
+
+CACHE_T = 2
+
+
+class Cosmos25CausalConv3d(nn.Conv3d):
+    """
+    Official-like causal 3D convolution.
+
+    Matches `CausalConv3d` in the official tokenizer: uses explicit F.pad and
+    supports a `cache_x` prefix for causal chunking.
+    """
+
+    def __init__(self, *args: Any, **kwargs: Any) -> None:
+        super().__init__(*args, **kwargs)
+        # padding order for F.pad: (W_left, W_right, H_left, H_right, T_left, T_right)
+        self._padding: tuple[int, ...] = (
+            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: torch.Tensor, cache_x: torch.Tensor | None = None) -> torch.Tensor:
+        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 Cosmos25RMSNorm(nn.Module):
+    """Official-like RMS_norm (uses learnable gamma and optional bias)."""
+
+    def __init__(self, dim: int, channel_first: bool = True, images: bool = True, bias: bool = False) -> None:
+        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 0.0
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        dim = 1 if self.channel_first else -1
+        return F.normalize(x, dim=dim) * self.scale * self.gamma + self.bias
+
+
+class Cosmos25Upsample(nn.Upsample):
+    """Official-like Upsample that is safe for bf16 (casts to fp32 internally)."""
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:  # type: ignore[override]
+        return super().forward(x.float()).type_as(x)
+
+
+class Cosmos25Resample(nn.Module):
+    """
+    Official-like Resample used for both spatial and temporal up/downsampling.
+    """
+
+    def __init__(self, dim: int, mode: str) -> None:
+        assert mode in ("none", "upsample2d", "upsample3d", "downsample2d", "downsample3d")
+        super().__init__()
+        self.dim = dim
+        self.mode = mode
+
+        if mode == "upsample2d":
+            self.resample = nn.Sequential(
+                Cosmos25Upsample(scale_factor=(2.0, 2.0), mode="nearest-exact"),
+                nn.Conv2d(dim, dim // 2, 3, padding=1),
+            )
+        elif mode == "upsample3d":
+            self.resample = nn.Sequential(
+                Cosmos25Upsample(scale_factor=(2.0, 2.0), mode="nearest-exact"),
+                nn.Conv2d(dim, dim // 2, 3, padding=1),
+            )
+            self.time_conv = Cosmos25CausalConv3d(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 = Cosmos25CausalConv3d(dim, dim, (3, 1, 1), stride=(2, 1, 1), padding=(0, 0, 0))
+        else:
+            self.resample = nn.Identity()
+
+    def forward(self, x: torch.Tensor, feat_cache: list[Any] | None = None, feat_idx: list[int] = [0]) -> torch.Tensor:
+        b, c, t, h, w = x.size()
+
+        # Temporal upsample uses a time-conv and then interleaves frames.
+        if self.mode == "upsample3d" and 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_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)
+
+        # Temporal downsample: time_conv consumes last-frame cache.
+        if self.mode == "downsample3d" and 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()
+                x = self.time_conv(torch.cat([feat_cache[idx][:, :, -1:, :, :], x], 2))
+                feat_cache[idx] = cache_x
+                feat_idx[0] += 1
+
+        return x
+
+
+class Cosmos25ResidualBlock(nn.Module):
+    def __init__(self, in_dim: int, out_dim: int, dropout: float = 0.0) -> None:
+        super().__init__()
+        self.in_dim = in_dim
+        self.out_dim = out_dim
+        self.residual = nn.Sequential(
+            Cosmos25RMSNorm(in_dim, images=False),
+            nn.SiLU(),
+            Cosmos25CausalConv3d(in_dim, out_dim, 3, padding=1),
+            Cosmos25RMSNorm(out_dim, images=False),
+            nn.SiLU(),
+            nn.Dropout(dropout),
+            Cosmos25CausalConv3d(out_dim, out_dim, 3, padding=1),
+        )
+        self.shortcut = Cosmos25CausalConv3d(in_dim, out_dim, 1) if in_dim != out_dim else nn.Identity()
+
+    def forward(self, x: torch.Tensor, feat_cache: list[Any] | None = None, feat_idx: list[int] = [0]) -> torch.Tensor:
+        h = self.shortcut(x)
+        for layer in self.residual:
+            if isinstance(layer, Cosmos25CausalConv3d) 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_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 Cosmos25AttentionBlock(nn.Module):
+    """Official-like causal self-attention with a single head."""
+
+    def __init__(self, dim: int) -> None:
+        super().__init__()
+        self.dim = dim
+        self.norm = Cosmos25RMSNorm(dim)
+        self.to_qkv = nn.Conv2d(dim, dim * 3, 1)
+        self.proj = nn.Conv2d(dim, dim, 1)
+        nn.init.zeros_(self.proj.weight)
+
+    def forward(self, x: torch.Tensor) -> torch.Tensor:
+        identity = x
+        b, c, t, h, w = x.size()
+        x2 = rearrange(x, "b c t h w -> (b t) c h w")
+        x2 = self.norm(x2)
+        q, k, v = (
+            self.to_qkv(x2)
+            .reshape(b * t, 1, c * 3, -1)
+            .permute(0, 1, 3, 2)
+            .contiguous()
+            .chunk(3, dim=-1)
+        )
+        x2 = F.scaled_dot_product_attention(q, k, v)
+        x2 = x2.squeeze(1).permute(0, 2, 1).reshape(b * t, c, h, w)
+        x2 = self.proj(x2)
+        x2 = rearrange(x2, "(b t) c h w-> b c t h w", t=t)
+        return x2 + identity
+
+
+class Cosmos25Encoder3d(nn.Module):
+    def __init__(
+        self,
+        dim: int = 96,
+        z_dim: int = 32,
+        dim_mult: list[int] = [1, 2, 4, 4],
+        num_res_blocks: int = 2,
+        attn_scales: list[float] = [],
+        temperal_downsample: list[bool] = [False, True, True],
+        dropout: float = 0.0,
+    ) -> None:
+        super().__init__()
+        dims = [dim * u for u in [1] + dim_mult]
+        scale = 1.0
+
+        self.conv1 = Cosmos25CausalConv3d(3, dims[0], 3, padding=1)
+
+        downsamples: list[nn.Module] = []
+        for i, (in_dim, out_dim) in enumerate(zip(dims[:-1], dims[1:])):
+            for _ in range(num_res_blocks):
+                downsamples.append(Cosmos25ResidualBlock(in_dim, out_dim, dropout))
+                if scale in attn_scales:
+                    downsamples.append(Cosmos25AttentionBlock(out_dim))
+                in_dim = out_dim
+
+            if i != len(dim_mult) - 1:
+                mode = "downsample3d" if temperal_downsample[i] else "downsample2d"
+                downsamples.append(Cosmos25Resample(out_dim, mode=mode))
+                scale /= 2.0
+        self.downsamples = nn.Sequential(*downsamples)
+
+        self.middle = nn.Sequential(
+            Cosmos25ResidualBlock(out_dim, out_dim, dropout),
+            Cosmos25AttentionBlock(out_dim),
+            Cosmos25ResidualBlock(out_dim, out_dim, dropout),
+        )
+
+        self.head = nn.Sequential(
+            Cosmos25RMSNorm(out_dim, images=False),
+            nn.SiLU(),
+            Cosmos25CausalConv3d(out_dim, z_dim, 3, padding=1),
+        )
+
+    def forward(self, x: torch.Tensor, feat_cache: list[Any] | None = None, feat_idx: list[int] = [0]) -> torch.Tensor:
+        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_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)
+
+        for layer in self.downsamples:
+            if feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)  # type: ignore[misc]
+            else:
+                x = layer(x)  # type: ignore[misc]
+
+        for layer in self.middle:
+            if isinstance(layer, Cosmos25ResidualBlock) and feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)
+            else:
+                x = layer(x)  # type: ignore[misc]
+
+        for layer in self.head:
+            if isinstance(layer, Cosmos25CausalConv3d) 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_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)  # type: ignore[misc]
+        return x
+
+
+class Cosmos25Decoder3d(nn.Module):
+    def __init__(
+        self,
+        dim: int = 96,
+        z_dim: int = 16,
+        dim_mult: list[int] = [1, 2, 4, 4],
+        num_res_blocks: int = 2,
+        attn_scales: list[float] = [],
+        temperal_upsample: list[bool] = [False, True, True],
+        dropout: float = 0.0,
+    ) -> None:
+        super().__init__()
+        dims = [dim * u for u in [dim_mult[-1]] + dim_mult[::-1]]
+        scale = 1.0 / 2 ** (len(dim_mult) - 2)
+
+        self.conv1 = Cosmos25CausalConv3d(z_dim, dims[0], 3, padding=1)
+        self.middle = nn.Sequential(
+            Cosmos25ResidualBlock(dims[0], dims[0], dropout),
+            Cosmos25AttentionBlock(dims[0]),
+            Cosmos25ResidualBlock(dims[0], dims[0], dropout),
+        )
+
+        upsamples: list[nn.Module] = []
+        for i, (in_dim, out_dim) in enumerate(zip(dims[:-1], dims[1:])):
+            if i in (1, 2, 3):
+                in_dim = in_dim // 2
+            for _ in range(num_res_blocks + 1):
+                upsamples.append(Cosmos25ResidualBlock(in_dim, out_dim, dropout))
+                if scale in attn_scales:
+                    upsamples.append(Cosmos25AttentionBlock(out_dim))
+                in_dim = out_dim
+            if i != len(dim_mult) - 1:
+                mode = "upsample3d" if temperal_upsample[i] else "upsample2d"
+                upsamples.append(Cosmos25Resample(out_dim, mode=mode))
+                scale *= 2.0
+        self.upsamples = nn.Sequential(*upsamples)
+
+        self.head = nn.Sequential(
+            Cosmos25RMSNorm(out_dim, images=False),
+            nn.SiLU(),
+            Cosmos25CausalConv3d(out_dim, 3, 3, padding=1),
+        )
+
+    def forward(self, x: torch.Tensor, feat_cache: list[Any] | None = None, feat_idx: list[int] = [0]) -> torch.Tensor:
+        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_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)
+
+        for layer in self.middle:
+            if isinstance(layer, Cosmos25ResidualBlock) and feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)
+            else:
+                x = layer(x)  # type: ignore[misc]
+
+        for layer in self.upsamples:
+            if feat_cache is not None:
+                x = layer(x, feat_cache, feat_idx)  # type: ignore[misc]
+            else:
+                x = layer(x)  # type: ignore[misc]
+
+        for layer in self.head:
+            if isinstance(layer, Cosmos25CausalConv3d) 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_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)  # type: ignore[misc]
+        return x
+
+
+def _count_cosmos25_conv3d(model: nn.Module) -> int:
+    return sum(1 for m in model.modules() if isinstance(m, Cosmos25CausalConv3d))
+
+
+class Cosmos25WanVAE(nn.Module):
+    """
+    A FastVideo-native copy of the *official-like* Wan2.1 VAE core.
+
+    Key properties:
+    - Module naming matches official tokenizer (`encoder`, `decoder`, `conv1`, `conv2`)
+      so it can consume `tokenizer.pth` keys directly.
+    - `encode()` returns **normalized** latents and `decode()` expects **normalized**
+      latents (Cosmos2.5 contract), matching `Wan2pt1VAEInterface`.
+    """
+
+    handles_latent_norm: bool = True
+    handles_latent_denorm: bool = True
+
+    def __init__(
+        self,
+        *,
+        device: torch.device | str = "cpu",
+        dtype: torch.dtype = torch.float32,
+        temporal_window: int = 4,
+        latents_mean: Optional[torch.Tensor] = None,
+        latents_std: Optional[torch.Tensor] = None,
+    ) -> None:
+        super().__init__()
+
+        # Official hyperparams for Cosmos2.5 tokenizer (Wan2.1 VAE).
+        cfg = dict(
+            dim=96,
+            z_dim=16,
+            dim_mult=[1, 2, 4, 4],
+            num_res_blocks=2,
+            attn_scales=[],
+            temperal_downsample=[False, True, True],
+            dropout=0.0,
+            temporal_window=temporal_window,
+        )
+        self.z_dim = 16
+        self.temporal_window = temporal_window
+
+        self.encoder = Cosmos25Encoder3d(
+            dim=cfg["dim"],
+            z_dim=cfg["z_dim"] * 2,
+            dim_mult=cfg["dim_mult"],
+            num_res_blocks=cfg["num_res_blocks"],
+            attn_scales=cfg["attn_scales"],
+            temperal_downsample=cfg["temperal_downsample"],
+            dropout=cfg["dropout"],
+        )
+        self.conv1 = Cosmos25CausalConv3d(self.z_dim * 2, self.z_dim * 2, 1)
+        self.conv2 = Cosmos25CausalConv3d(self.z_dim, self.z_dim, 1)
+        self.decoder = Cosmos25Decoder3d(
+            dim=cfg["dim"],
+            z_dim=cfg["z_dim"],
+            dim_mult=cfg["dim_mult"],
+            num_res_blocks=cfg["num_res_blocks"],
+            attn_scales=cfg["attn_scales"],
+            temperal_upsample=list(cfg["temperal_downsample"])[::-1],
+            dropout=cfg["dropout"],
+        )
+
+        # Default Cosmos2.5 latent stats (shared with configs).
+        if latents_mean is None:
+            latents_mean = torch.tensor(
+                [
+                    -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,
+                ],
+                dtype=torch.float32,
+            ).view(1, 16, 1, 1, 1)
+        if latents_std is None:
+            latents_std = torch.tensor(
+                [
+                    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,
+                ],
+                dtype=torch.float32,
+            ).view(1, 16, 1, 1, 1)
+
+        self.register_buffer("_latents_mean", latents_mean, persistent=False)
+        self.register_buffer("_latents_std", latents_std, persistent=False)
+
+        self.to(device=device, dtype=dtype)
+        self.clear_cache()
+
+    def clear_cache(self) -> None:
+        # Decoder cache
+        self._conv_num = _count_cosmos25_conv3d(self.decoder)
+        self._conv_idx = [0]
+        self._feat_map: list[Any] = [None] * self._conv_num
+        # Encoder cache
+        self._enc_conv_num = _count_cosmos25_conv3d(self.encoder)
+        self._enc_conv_idx = [0]
+        self._enc_feat_map: list[Any] = [None] * self._enc_conv_num
+
+    def _scale(self, like: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+        mean = self._latents_mean.to(device=like.device, dtype=like.dtype)
+        std = self._latents_std.to(device=like.device, dtype=like.dtype)
+        return mean, 1.0 / std
+
+    def _i0_encode(self, x: torch.Tensor) -> torch.Tensor:
+        return self.encoder(x[:, :, :1, :, :], feat_cache=self._enc_feat_map, feat_idx=self._enc_conv_idx)
+
+    def _i0_decode(self, x: torch.Tensor) -> torch.Tensor:
+        return self.decoder(x[:, :, 0:1, :, :], feat_cache=self._feat_map, feat_idx=self._conv_idx)
+
+    def encode(self, x: torch.Tensor) -> _TensorLatentDist:
+        """
+        Encode to *normalized* latents (Cosmos contract).
+        """
+        self.clear_cache()
+        t = x.shape[2]
+        iters = 1 + (t - 1) // self.temporal_window
+
+        for i in range(iters):
+            self._enc_conv_idx = [0]
+            if i == 0:
+                out = self._i0_encode(x)
+            else:
+                out_ = self.encoder(
+                    x[:, :, 1 + self.temporal_window * (i - 1) : 1 + self.temporal_window * i, :, :],
+                    feat_cache=self._enc_feat_map,
+                    feat_idx=self._enc_conv_idx,
+                )
+                out = torch.cat([out, out_], 2)
+
+        if (t - 1) % self.temporal_window:
+            self._enc_conv_idx = [0]
+            out_ = self.encoder(
+                x[:, :, 1 + self.temporal_window * (iters - 1) :, :, :],
+                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)
+        mean, inv_std = self._scale(mu)
+        z_norm = (mu - mean) * inv_std
+        self.clear_cache()
+        return _TensorLatentDist(z_norm)
+
+    def decode(self, latent: torch.Tensor) -> torch.Tensor:
+        """
+        Decode from *normalized* latents (Cosmos contract).
+        """
+        self.clear_cache()
+        mean, inv_std = self._scale(latent)
+        z = latent / inv_std + mean  # z = z_norm * std + mean
+
+        iter_ = z.shape[2]
+        x = self.conv2(z)
+        for i in range(iter_):
+            self._conv_idx = [0]
+            if i == 0:
+                out = self._i0_decode(x)
+            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
+
+    # --- Interface helpers (match official Wan2pt1VAEInterface) ---
+    def get_latent_num_frames(self, num_pixel_frames: int) -> int:
+        return 1 + (int(num_pixel_frames) - 1) // 4
+
+    def get_pixel_num_frames(self, num_latent_frames: int) -> int:
+        return (int(num_latent_frames) - 1) * 4 + 1
+
+    @property
+    def spatial_compression_factor(self) -> int:
+        return 8
+
+    @property
+    def temporal_compression_factor(self) -> int:
+        return 4
+
+    @property
+    def latent_ch(self) -> int:
+        return 16
+
+
+
+
+
+
+
+
+

fastvideo/pipelines/basic/cosmos/cosmos2_5_pipeline.py

@@ -0,0 +1,61 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Cosmos 2.5 pipeline entry (staged pipeline)."""
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.pipelines.composed_pipeline_base import ComposedPipelineBase
+from fastvideo.pipelines.stages import (ConditioningStage,
+                                        Cosmos25DenoisingStage,
+                                        Cosmos25LatentPreparationStage,
+                                        DecodingStage, InputValidationStage,
+                                        Cosmos25TextEncodingStage,
+                                        Cosmos25TimestepPreparationStage)
+
+logger = init_logger(__name__)
+
+
+class Cosmos2_5Pipeline(ComposedPipelineBase):
+    """Cosmos 2.5 video generation pipeline."""
+
+    _required_config_modules = [
+        "text_encoder", "tokenizer", "vae", "transformer", "scheduler",
+        "safety_checker"
+    ]
+
+    def create_pipeline_stages(self, fastvideo_args: FastVideoArgs):
+        logger.info("Creating Cosmos 2.5 pipeline stages...")
+
+        self.add_stage(stage_name="input_validation_stage",
+                       stage=InputValidationStage())
+
+        self.add_stage(
+            stage_name="prompt_encoding_stage",
+            stage=Cosmos25TextEncodingStage(
+                text_encoder=self.get_module("text_encoder"), ),
+        )
+
+        self.add_stage(stage_name="conditioning_stage",
+                       stage=ConditioningStage())
+
+        self.add_stage(stage_name="timestep_preparation_stage",
+                       stage=Cosmos25TimestepPreparationStage(
+                           scheduler=self.get_module("scheduler")))
+
+        self.add_stage(stage_name="latent_preparation_stage",
+                       stage=Cosmos25LatentPreparationStage(
+                           scheduler=self.get_module("scheduler"),
+                           transformer=self.get_module("transformer"),
+                           vae=self.get_module("vae")))
+
+        self.add_stage(stage_name="denoising_stage",
+                       stage=Cosmos25DenoisingStage(
+                           transformer=self.get_module("transformer"),
+                           scheduler=self.get_module("scheduler")))
+
+        self.add_stage(stage_name="decoding_stage",
+                       stage=DecodingStage(vae=self.get_module("vae")))
+        logger.info("Cosmos 2.5 pipeline stages created")
+
+
+# Entry point for pipeline registry
+EntryClass = Cosmos2_5Pipeline

fastvideo/pipelines/basic/longcat/__init__.py

@@ -2,5 +2,10 @@
 """LongCat pipeline module."""
 
 from fastvideo.pipelines.basic.longcat.longcat_pipeline import LongCatPipeline
+from fastvideo.pipelines.basic.longcat.longcat_i2v_pipeline import LongCatImageToVideoPipeline
+from fastvideo.pipelines.basic.longcat.longcat_vc_pipeline import LongCatVideoContinuationPipeline
 
-__all__ = ["LongCatPipeline"]
+__all__ = [
+    "LongCatPipeline", "LongCatImageToVideoPipeline",
+    "LongCatVideoContinuationPipeline"
+]

fastvideo/pipelines/basic/longcat/longcat_i2v_pipeline.py

@@ -0,0 +1,148 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat Image-to-Video pipeline implementation.
+
+This module implements I2V (Image-to-Video) generation for LongCat using Tier 3
+conditioning with timestep masking, num_cond_latents support, and RoPE skipping.
+
+Supports:
+- Basic I2V (50 steps, guidance_scale=4.0)
+- Distilled I2V with LoRA (16 steps, guidance_scale=1.0)
+- Refinement I2V for 720p upscaling (with refinement LoRA + BSA)
+"""
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.pipelines import ComposedPipelineBase, LoRAPipeline
+from fastvideo.pipelines.stages import (
+    DecodingStage,
+    InputValidationStage,
+    TextEncodingStage,
+    TimestepPreparationStage,
+)
+from fastvideo.pipelines.stages.longcat_image_vae_encoding import LongCatImageVAEEncodingStage
+from fastvideo.pipelines.stages.longcat_i2v_latent_preparation import LongCatI2VLatentPreparationStage
+from fastvideo.pipelines.stages.longcat_i2v_denoising import LongCatI2VDenoisingStage
+from fastvideo.pipelines.stages.longcat_refine_init import LongCatRefineInitStage
+from fastvideo.pipelines.stages.longcat_refine_timestep import LongCatRefineTimestepStage
+
+logger = init_logger(__name__)
+
+
+class LongCatImageToVideoPipeline(LoRAPipeline, ComposedPipelineBase):
+    """
+    LongCat Image-to-Video pipeline.
+    
+    Generates video from a single input image using Tier 3 I2V conditioning:
+    - Per-frame timestep masking (timestep[:, 0] = 0)
+    - num_cond_latents parameter to transformer
+    - RoPE skipping for conditioning frames
+    - Selective denoising (skip first frame in scheduler)
+    """
+
+    _required_config_modules = [
+        "text_encoder", "tokenizer", "vae", "transformer", "scheduler"
+    ]
+
+    def initialize_pipeline(self, fastvideo_args: FastVideoArgs):
+        """Initialize LongCat-specific components."""
+        # Same BSA initialization as base LongCat pipeline
+        pipeline_config = fastvideo_args.pipeline_config
+        transformer = self.get_module("transformer", None)
+        if transformer is None:
+            return
+
+        # Enable BSA if configured
+        if pipeline_config.enable_bsa:
+            bsa_params_cfg = getattr(pipeline_config, 'bsa_params', None) or {}
+            sparsity = getattr(pipeline_config, 'bsa_sparsity', None)
+            cdf_threshold = getattr(pipeline_config, 'bsa_cdf_threshold', None)
+            chunk_q = getattr(pipeline_config, 'bsa_chunk_q', None)
+            chunk_k = getattr(pipeline_config, 'bsa_chunk_k', None)
+
+            effective_bsa_params = dict(bsa_params_cfg) if isinstance(
+                bsa_params_cfg, dict) else {}
+            if sparsity is not None:
+                effective_bsa_params['sparsity'] = sparsity
+            if cdf_threshold is not None:
+                effective_bsa_params['cdf_threshold'] = cdf_threshold
+            if chunk_q is not None:
+                effective_bsa_params['chunk_3d_shape_q'] = chunk_q
+            if chunk_k is not None:
+                effective_bsa_params['chunk_3d_shape_k'] = chunk_k
+
+            # Provide defaults
+            effective_bsa_params.setdefault('sparsity', 0.9375)
+            effective_bsa_params.setdefault('chunk_3d_shape_q', [4, 4, 4])
+            effective_bsa_params.setdefault('chunk_3d_shape_k', [4, 4, 4])
+
+            if hasattr(transformer, 'enable_bsa'):
+                logger.info("Enabling BSA for LongCat I2V transformer")
+                transformer.enable_bsa()
+                if hasattr(transformer, 'blocks'):
+                    try:
+                        for blk in transformer.blocks:
+                            if hasattr(blk, 'self_attn'):
+                                blk.self_attn.bsa_params = effective_bsa_params
+                    except Exception as e:
+                        logger.warning("Failed to set BSA params: %s", e)
+                logger.info("BSA parameters: %s", effective_bsa_params)
+        else:
+            if hasattr(transformer, 'disable_bsa'):
+                transformer.disable_bsa()
+
+    def create_pipeline_stages(self, fastvideo_args: FastVideoArgs):
+        """Set up I2V-specific pipeline stages."""
+
+        # 1. Input validation
+        self.add_stage(stage_name="input_validation_stage",
+                       stage=InputValidationStage())
+
+        # 2. Text encoding (same as T2V)
+        self.add_stage(stage_name="prompt_encoding_stage",
+                       stage=TextEncodingStage(
+                           text_encoders=[self.get_module("text_encoder")],
+                           tokenizers=[self.get_module("tokenizer")],
+                       ))
+
+        # 3. Image VAE encoding (for I2V - skipped in refinement mode)
+        self.add_stage(
+            stage_name="image_vae_encoding_stage",
+            stage=LongCatImageVAEEncodingStage(vae=self.get_module("vae")))
+
+        # 4. Refinement initialization (skipped if not refining)
+        self.add_stage(stage_name="longcat_refine_init_stage",
+                       stage=LongCatRefineInitStage(vae=self.get_module("vae")))
+
+        # 5. Timestep preparation (generic)
+        self.add_stage(stage_name="timestep_preparation_stage",
+                       stage=TimestepPreparationStage(
+                           scheduler=self.get_module("scheduler")))
+
+        # 6. Refinement timestep override (skipped if not refining)
+        self.add_stage(stage_name="longcat_refine_timestep_stage",
+                       stage=LongCatRefineTimestepStage(
+                           scheduler=self.get_module("scheduler")))
+
+        # 7. Latent preparation with I2V conditioning
+        self.add_stage(stage_name="latent_preparation_stage",
+                       stage=LongCatI2VLatentPreparationStage(
+                           scheduler=self.get_module("scheduler"),
+                           transformer=self.get_module("transformer")))
+
+        # 8. Denoising with I2V support
+        self.add_stage(stage_name="denoising_stage",
+                       stage=LongCatI2VDenoisingStage(
+                           transformer=self.get_module("transformer"),
+                           transformer_2=self.get_module("transformer_2", None),
+                           scheduler=self.get_module("scheduler"),
+                           vae=self.get_module("vae"),
+                           pipeline=self))
+
+        # 9. Decoding
+        self.add_stage(stage_name="decoding_stage",
+                       stage=DecodingStage(vae=self.get_module("vae"),
+                                           pipeline=self))
+
+
+EntryClass = LongCatImageToVideoPipeline

fastvideo/pipelines/basic/longcat/longcat_pipeline.py

@@ -1,9 +1,9 @@
 # SPDX-License-Identifier: Apache-2.0
 """
-LongCat video diffusion pipeline implementation (Phase 1: Wrapper).
+LongCat video diffusion pipeline implementation.
 
-This module contains a wrapper implementation of the LongCat video diffusion pipeline
-using FastVideo's modular pipeline architecture with the original LongCat modules.
+This module implements the LongCat video diffusion pipeline using FastVideo's
+modular pipeline architecture.
 """
 
 from fastvideo.fastvideo_args import FastVideoArgs
@@ -26,9 +26,6 @@ logger = init_logger(__name__)
 class LongCatPipeline(LoRAPipeline, ComposedPipelineBase):
     """
     LongCat video diffusion pipeline with LoRA support.
-    
-    Phase 1 implementation using wrapper modules from third_party/longcat_video.
-    This validates the pipeline infrastructure before full FastVideo integration.
     """
 
     _required_config_modules = [

fastvideo/pipelines/basic/longcat/longcat_vc_pipeline.py

@@ -0,0 +1,168 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat Video Continuation (VC) pipeline implementation.
+
+This module implements VC (Video Continuation) generation for LongCat with
+KV cache optimization for 2-3x speedup.
+
+Supports:
+- Basic VC (50 steps, guidance_scale=4.0)
+- Distilled VC with LoRA (16 steps, guidance_scale=1.0)
+- KV cache for conditioning frames
+"""
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.pipelines import ComposedPipelineBase, LoRAPipeline
+from fastvideo.pipelines.stages import (
+    DecodingStage,
+    InputValidationStage,
+    TextEncodingStage,
+    TimestepPreparationStage,
+)
+from fastvideo.pipelines.stages.longcat_video_vae_encoding import LongCatVideoVAEEncodingStage
+from fastvideo.pipelines.stages.longcat_i2v_latent_preparation import LongCatI2VLatentPreparationStage
+from fastvideo.pipelines.stages.longcat_kv_cache_init import LongCatKVCacheInitStage
+from fastvideo.pipelines.stages.longcat_vc_denoising import LongCatVCDenoisingStage
+
+logger = init_logger(__name__)
+
+
+class LongCatVideoContinuationPipeline(LoRAPipeline, ComposedPipelineBase):
+    """
+    LongCat Video Continuation pipeline.
+    
+    Generates video continuation from multiple conditioning frames using
+    optional KV cache for 2-3x speedup.
+    
+    Key features:
+    - Takes video input (13+ frames typically)
+    - Encodes conditioning frames via VAE
+    - Optionally pre-computes KV cache for conditioning
+    - Uses cached K/V during denoising for speedup
+    - Concatenates conditioning back after denoising
+    """
+
+    _required_config_modules = [
+        "text_encoder", "tokenizer", "vae", "transformer", "scheduler"
+    ]
+
+    def initialize_pipeline(self, fastvideo_args: FastVideoArgs):
+        """Initialize LongCat-specific components."""
+        pipeline_config = fastvideo_args.pipeline_config
+        transformer = self.get_module("transformer", None)
+        if transformer is None:
+            return
+
+        # Enable BSA if configured (for VC, BSA may not be needed)
+        if getattr(pipeline_config, 'enable_bsa', False):
+            bsa_params_cfg = getattr(pipeline_config, 'bsa_params', None) or {}
+            sparsity = getattr(pipeline_config, 'bsa_sparsity', None)
+            cdf_threshold = getattr(pipeline_config, 'bsa_cdf_threshold', None)
+            chunk_q = getattr(pipeline_config, 'bsa_chunk_q', None)
+            chunk_k = getattr(pipeline_config, 'bsa_chunk_k', None)
+
+            effective_bsa_params = dict(bsa_params_cfg) if isinstance(
+                bsa_params_cfg, dict) else {}
+            if sparsity is not None:
+                effective_bsa_params['sparsity'] = sparsity
+            if cdf_threshold is not None:
+                effective_bsa_params['cdf_threshold'] = cdf_threshold
+            if chunk_q is not None:
+                effective_bsa_params['chunk_3d_shape_q'] = chunk_q
+            if chunk_k is not None:
+                effective_bsa_params['chunk_3d_shape_k'] = chunk_k
+
+            # Provide defaults
+            effective_bsa_params.setdefault('sparsity', 0.9375)
+            effective_bsa_params.setdefault('chunk_3d_shape_q', [4, 4, 4])
+            effective_bsa_params.setdefault('chunk_3d_shape_k', [4, 4, 4])
+
+            if hasattr(transformer, 'enable_bsa'):
+                logger.info("Enabling BSA for LongCat VC transformer")
+                transformer.enable_bsa()
+                if hasattr(transformer, 'blocks'):
+                    try:
+                        for blk in transformer.blocks:
+                            if hasattr(blk, 'self_attn'):
+                                blk.self_attn.bsa_params = effective_bsa_params
+                    except Exception as e:
+                        logger.warning("Failed to set BSA params: %s", e)
+                logger.info("BSA parameters: %s", effective_bsa_params)
+        else:
+            if hasattr(transformer, 'disable_bsa'):
+                transformer.disable_bsa()
+
+    def create_pipeline_stages(self, fastvideo_args: FastVideoArgs):
+        """Set up VC-specific pipeline stages."""
+
+        # 1. Input validation
+        self.add_stage(stage_name="input_validation_stage",
+                       stage=InputValidationStage())
+
+        # 2. Text encoding
+        self.add_stage(stage_name="prompt_encoding_stage",
+                       stage=TextEncodingStage(
+                           text_encoders=[self.get_module("text_encoder")],
+                           tokenizers=[self.get_module("tokenizer")],
+                       ))
+
+        # 3. Video VAE encoding (encodes conditioning frames)
+        self.add_stage(
+            stage_name="video_vae_encoding_stage",
+            stage=LongCatVideoVAEEncodingStage(vae=self.get_module("vae")))
+
+        # 4. Timestep preparation
+        self.add_stage(stage_name="timestep_preparation_stage",
+                       stage=TimestepPreparationStage(
+                           scheduler=self.get_module("scheduler")))
+
+        # 5. Latent preparation (reuse I2V stage - it handles video_latent too)
+        self.add_stage(stage_name="latent_preparation_stage",
+                       stage=LongCatVCLatentPreparationStage(
+                           scheduler=self.get_module("scheduler"),
+                           transformer=self.get_module("transformer")))
+
+        # 6. KV cache initialization (optional, based on config)
+        # This is always added but will skip if use_kv_cache=False
+        self.add_stage(stage_name="kv_cache_init_stage",
+                       stage=LongCatKVCacheInitStage(
+                           transformer=self.get_module("transformer")))
+
+        # 7. Denoising with VC and KV cache support
+        self.add_stage(stage_name="denoising_stage",
+                       stage=LongCatVCDenoisingStage(
+                           transformer=self.get_module("transformer"),
+                           transformer_2=self.get_module("transformer_2", None),
+                           scheduler=self.get_module("scheduler"),
+                           vae=self.get_module("vae"),
+                           pipeline=self))
+
+        # 8. Decoding
+        self.add_stage(stage_name="decoding_stage",
+                       stage=DecodingStage(vae=self.get_module("vae"),
+                                           pipeline=self))
+
+
+class LongCatVCLatentPreparationStage(LongCatI2VLatentPreparationStage):
+    """
+    Prepare latents with video conditioning for first N frames.
+    
+    Extends I2V latent preparation to handle video_latent (multiple frames)
+    instead of image_latent (single frame).
+    """
+
+    def forward(self, batch, fastvideo_args):
+        """Prepare latents with VC conditioning."""
+
+        # Check if we have video_latent (from VC encoding stage)
+        video_latent = getattr(batch, 'video_latent', None)
+        if video_latent is not None:
+            # Set image_latent to video_latent for parent class compatibility
+            batch.image_latent = video_latent
+
+        # Call parent class forward
+        return super().forward(batch, fastvideo_args)
+
+
+EntryClass = LongCatVideoContinuationPipeline

fastvideo/pipelines/basic/turbodiffusion/__init__.py

@@ -0,0 +1,6 @@
+from fastvideo.pipelines.basic.turbodiffusion.turbodiffusion_pipeline import (
+    TurboDiffusionPipeline, )
+from fastvideo.pipelines.basic.turbodiffusion.turbodiffusion_i2v_pipeline import (
+    TurboDiffusionI2VPipeline, )
+
+__all__ = ["TurboDiffusionPipeline", "TurboDiffusionI2VPipeline"]

fastvideo/pipelines/basic/turbodiffusion/turbodiffusion_i2v_pipeline.py

@@ -0,0 +1,89 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+TurboDiffusion I2V (Image-to-Video) Pipeline Implementation.
+
+This module contains an implementation of the TurboDiffusion I2V pipeline
+for 1-4 step image-to-video generation using rCM (recurrent Consistency Model)
+sampling with SLA (Sparse-Linear Attention).
+
+Key differences from T2V:
+- Uses dual models (high/low noise) with boundary switching
+- sigma_max=200 (vs 80 for T2V)
+- Mask conditioning with encoded first frame
+"""
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.schedulers.scheduling_rcm import RCMScheduler
+from fastvideo.pipelines import ComposedPipelineBase, LoRAPipeline
+from fastvideo.pipelines.stages import (ConditioningStage, DecodingStage,
+                                        DenoisingStage, ImageVAEEncodingStage,
+                                        InputValidationStage,
+                                        LatentPreparationStage,
+                                        TextEncodingStage,
+                                        TimestepPreparationStage)
+
+logger = init_logger(__name__)
+
+
+class TurboDiffusionI2VPipeline(LoRAPipeline, ComposedPipelineBase):
+    """
+    TurboDiffusion I2V pipeline for 1-4 step image-to-video generation.
+
+    Uses RCM scheduler, SLA attention, and dual model switching for
+    high-quality I2V generation.
+    """
+
+    _required_config_modules = [
+        "text_encoder", "tokenizer", "vae", "transformer", "transformer_2",
+        "scheduler"
+    ]
+
+    def initialize_pipeline(self, fastvideo_args: FastVideoArgs):
+        # Use RCM scheduler with higher sigma_max for I2V
+        logger.info(
+            "Initializing RCM scheduler for TurboDiffusion I2V (sigma_max=200)")
+        self.modules["scheduler"] = RCMScheduler(sigma_max=200.0)
+
+    def create_pipeline_stages(self, fastvideo_args: FastVideoArgs) -> None:
+        """Set up pipeline stages with proper dependency injection."""
+
+        self.add_stage(stage_name="input_validation_stage",
+                       stage=InputValidationStage())
+
+        self.add_stage(stage_name="prompt_encoding_stage",
+                       stage=TextEncodingStage(
+                           text_encoders=[self.get_module("text_encoder")],
+                           tokenizers=[self.get_module("tokenizer")],
+                       ))
+
+        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"),
+                           transformer=self.get_module("transformer", None)))
+
+        # I2V: Encode initial image to latent space
+        self.add_stage(stage_name="image_latent_preparation_stage",
+                       stage=ImageVAEEncodingStage(vae=self.get_module("vae")))
+
+        self.add_stage(stage_name="denoising_stage",
+                       stage=DenoisingStage(
+                           transformer=self.get_module("transformer"),
+                           transformer_2=self.get_module("transformer_2", None),
+                           scheduler=self.get_module("scheduler"),
+                           vae=self.get_module("vae"),
+                           pipeline=self))
+
+        self.add_stage(stage_name="decoding_stage",
+                       stage=DecodingStage(vae=self.get_module("vae"),
+                                           pipeline=self))
+
+
+EntryClass = TurboDiffusionI2VPipeline

fastvideo/pipelines/basic/turbodiffusion/turbodiffusion_pipeline.py

@@ -36,11 +36,6 @@ class TurboDiffusionPipeline(LoRAPipeline, ComposedPipelineBase):
         logger.info("Initializing RCM scheduler for TurboDiffusion")
         self.modules["scheduler"] = RCMScheduler(sigma_max=80.0)
 
-        # Store checkpoint path for later loading
-        self._turbodiffusion_checkpoint = getattr(fastvideo_args,
-                                                  'turbodiffusion_checkpoint',
-                                                  None)
-
     def create_pipeline_stages(self, fastvideo_args: FastVideoArgs) -> None:
         """Set up pipeline stages with proper dependency injection."""
 

fastvideo/pipelines/composed_pipeline_base.py

@@ -201,8 +201,6 @@ class ComposedPipelineBase(ABC):
             # fwd, bwd, and other operations' precision.
             assert fastvideo_args.pipeline_config.dit_precision == 'fp32', 'only fp32 is supported for training'
 
-        logger.info("fastvideo_args in from_pretrained: %s", fastvideo_args)
-
         pipe = cls(model_path,
                    fastvideo_args,
                    required_config_modules=required_config_modules,

fastvideo/pipelines/lora_pipeline.py

@@ -1,7 +1,9 @@
 # SPDX-License-Identifier: Apache-2.0
 from collections import defaultdict
 from collections.abc import Hashable
+from contextlib import nullcontext
 from typing import Any
+from collections.abc import Generator
 
 import torch
 import torch.distributed as dist
@@ -12,8 +14,13 @@ from torch.distributed.tensor import DTensor
 
 from fastvideo.distributed import get_local_torch_device
 from fastvideo.fastvideo_args import FastVideoArgs, TrainingArgs
-from fastvideo.layers.lora.linear import (BaseLayerWithLoRA, get_lora_layer,
-                                          replace_submodule)
+from fastvideo.hooks.hooks import ModuleHookManager
+from fastvideo.hooks.layerwise_offload import LayerwiseOffloadHook
+from fastvideo.layers.lora.linear import (
+    BaseLayerWithLoRA,
+    get_lora_layer,
+    replace_submodule,
+)
 from fastvideo.logger import init_logger
 from fastvideo.models.loader.utils import get_param_names_mapping
 from fastvideo.pipelines.composed_pipeline_base import ComposedPipelineBase
@@ -22,17 +29,89 @@ from fastvideo.utils import maybe_download_lora
 logger = init_logger(__name__)
 
 
+def _get_hook_ctx(module: nn.Module | None):
+    if module is None:
+        return nullcontext()
+    hook_mgr = ModuleHookManager.get_from(module)
+    if hook_mgr is not None:
+        offload_hook = hook_mgr.forward_hooks.get(LayerwiseOffloadHook.name())
+        if offload_hook is not None:
+            return offload_hook.mutate_params_scope()  # type: ignore
+    return nullcontext()
+
+
+def _named_module_by_prefix(
+    module: nn.Module, prefixes: list[str]
+) -> list[tuple[str | None, list[tuple[str, nn.Module]]]]:
+    none_list: list[tuple[str, nn.Module]] = []
+    prefix_list: list[tuple[str, list[tuple[str, nn.Module]]]] = [
+        (prefix, []) for prefix in prefixes
+    ]
+    for name, submodule in module.named_modules():
+        for cur_prefix, cur_list in prefix_list:
+            # we should exclude e.g. block.1 and block.12.attn
+            if name.startswith(cur_prefix + "."):
+                cur_list.append((name, submodule))
+                break
+        else:
+            none_list.append((name, submodule))
+    return prefix_list + [(None, none_list)]  # type: ignore
+
+
+class LoRAModelLayers:
+
+    def __init__(self, block_list: list[tuple[str, nn.Module]]) -> None:
+        # block_name -> {layer_name -> layer}
+        self.block_to_lora_layers: dict[str, dict[str, BaseLayerWithLoRA]] = {}
+        # layer_name -> block_name
+        self.lora_layers_to_block: dict[str, str | None] = {}
+        self.other_lora_layers: dict[str, BaseLayerWithLoRA] = {}
+        self.block_mapping = dict(block_list)
+
+    def add_lora_layer(self, block_name: str | None, layer_name: str,
+                       layer: BaseLayerWithLoRA):
+        if block_name is None:
+            self.other_lora_layers[layer_name] = layer
+            self.lora_layers_to_block[layer_name] = None
+        else:
+            if block_name not in self.block_to_lora_layers:
+                self.block_to_lora_layers[block_name] = {}
+            self.block_to_lora_layers[block_name][layer_name] = layer
+            self.lora_layers_to_block[layer_name] = block_name
+
+    def all_lora_layers(
+        self, ) -> Generator[tuple[str, BaseLayerWithLoRA], Any, None]:
+        for block_layers in self.block_to_lora_layers.values():
+            for name, layer in block_layers.items():
+                yield name, layer
+        for name, layer in self.other_lora_layers.items():
+            yield name, layer
+
+    def lora_layers_by_block(
+        self,
+    ) -> Generator[
+            tuple[nn.Module | None, dict[str, BaseLayerWithLoRA]],
+            Any,
+            None,
+    ]:
+        for block_name, layers in self.block_to_lora_layers.items():
+            yield self.block_mapping[block_name], layers
+        yield None, self.other_lora_layers
+
+
 class LoRAPipeline(ComposedPipelineBase):
     """
     Pipeline that supports injecting LoRA adapters into the diffusion transformer.
     TODO: support training.
     """
+
     lora_adapters: dict[str, dict[str, torch.Tensor]] = defaultdict(
         dict
     )  # state dicts of loaded lora adapters (includes lora_A, lora_B, and lora_alpha)
     cur_adapter_name: str = ""
     cur_adapter_path: str = ""
-    lora_layers: dict[str, dict[str, BaseLayerWithLoRA]] = {}
+    # model_name -> layers
+    lora_layers: dict[str, LoRAModelLayers] = {}
     fastvideo_args: FastVideoArgs | TrainingArgs
     exclude_lora_layers: dict[str, list[str]] = {}
     device: torch.device = get_local_torch_device()
@@ -48,10 +127,10 @@ class LoRAPipeline(ComposedPipelineBase):
         self.device = get_local_torch_device()
         # build list of trainable transformers
         for transformer_name in self.trainable_transformer_names:
-            if transformer_name in self.modules and self.modules[
-                    transformer_name] is not None:
-                self.trainable_transformer_modules[
-                    transformer_name] = self.modules[transformer_name]
+            if (transformer_name in self.modules
+                    and self.modules[transformer_name] is not None):
+                self.trainable_transformer_modules[transformer_name] = (
+                    self.modules[transformer_name])
             # check for transformer_2 in case of Wan2.2 MoE or fake_score_transformer_2
             if transformer_name.endswith("_2"):
                 raise ValueError(
@@ -59,19 +138,23 @@ class LoRAPipeline(ComposedPipelineBase):
                 )
 
             secondary_transformer_name = transformer_name + "_2"
-            if secondary_transformer_name in self.modules and self.modules[
-                    secondary_transformer_name] is not None:
+            if (secondary_transformer_name in self.modules
+                    and self.modules[secondary_transformer_name] is not None):
                 self.trainable_transformer_modules[
                     secondary_transformer_name] = self.modules[
                         secondary_transformer_name]
 
-        logger.info("trainable_transformer_modules: %s",
-                    self.trainable_transformer_modules.keys())
+        logger.info(
+            "trainable_transformer_modules: %s",
+            self.trainable_transformer_modules.keys(),
+        )
 
-        for transformer_name, transformer_module in self.trainable_transformer_modules.items(
-        ):
-            self.exclude_lora_layers[
-                transformer_name] = transformer_module.config.arch_config.exclude_lora_layers
+        for (
+                transformer_name,
+                transformer_module,
+        ) in self.trainable_transformer_modules.items():
+            self.exclude_lora_layers[transformer_name] = (
+                transformer_module.config.arch_config.exclude_lora_layers)
         self.lora_target_modules = self.fastvideo_args.lora_target_modules
         self.lora_path = self.fastvideo_args.lora_path
         self.lora_nickname = self.fastvideo_args.lora_nickname
@@ -83,20 +166,33 @@ class LoRAPipeline(ComposedPipelineBase):
                 self.fastvideo_args.lora_alpha = self.fastvideo_args.lora_rank
             self.lora_rank = self.fastvideo_args.lora_rank  # type: ignore
             self.lora_alpha = self.fastvideo_args.lora_alpha  # type: ignore
-            logger.info("Using LoRA training with rank %d and alpha %d",
-                        self.lora_rank, self.lora_alpha)
+            logger.info(
+                "Using LoRA training with rank %d and alpha %d",
+                self.lora_rank,
+                self.lora_alpha,
+            )
             if self.lora_target_modules is None:
                 self.lora_target_modules = [
-                    "q_proj", "k_proj", "v_proj", "o_proj", "to_q", "to_k",
-                    "to_v", "to_out", "to_qkv", "to_gate_compress"
+                    "q_proj",
+                    "k_proj",
+                    "v_proj",
+                    "o_proj",
+                    "to_q",
+                    "to_k",
+                    "to_v",
+                    "to_out",
+                    "to_qkv",
+                    "to_gate_compress",
                 ]
                 logger.info(
                     "Using default lora_target_modules for all transformers: %s",
-                    self.lora_target_modules)
+                    self.lora_target_modules,
+                )
             else:
                 logger.warning(
                     "Using custom lora_target_modules for all transformers, which may not be intended: %s",
-                    self.lora_target_modules)
+                    self.lora_target_modules,
+                )
 
             self.convert_to_lora_layers()
         # Inference
@@ -104,7 +200,8 @@ class LoRAPipeline(ComposedPipelineBase):
             self.convert_to_lora_layers()
             self.set_lora_adapter(
                 self.lora_nickname,  # type: ignore
-                self.lora_path)  # type: ignore
+                self.lora_path,
+            )  # type: ignore
 
     def is_target_layer(self, module_name: str) -> bool:
         if self.lora_target_modules is None:
@@ -114,9 +211,9 @@ class LoRAPipeline(ComposedPipelineBase):
 
     def set_trainable(self) -> None:
 
-        def set_lora_grads(lora_layers: dict[str, BaseLayerWithLoRA],
+        def set_lora_grads(lora_layers: LoRAModelLayers,
                            device_mesh: DeviceMesh):
-            for name, layer in lora_layers.items():
+            for name, layer in lora_layers.all_lora_layers():
                 layer.lora_A.requires_grad_(True)
                 layer.lora_B.requires_grad_(True)
                 layer.base_layer.requires_grad_(False)
@@ -131,10 +228,15 @@ class LoRAPipeline(ComposedPipelineBase):
             super().set_trainable()
             return
 
-        device_mesh = init_device_mesh("cuda", (dist.get_world_size(), 1),
-                                       mesh_dim_names=["fake", "replicate"])
-        for transformer_name, transformer_module in self.trainable_transformer_modules.items(
-        ):
+        device_mesh = init_device_mesh(
+            "cuda",
+            (dist.get_world_size(), 1),
+            mesh_dim_names=["fake", "replicate"],
+        )
+        for (
+                transformer_name,
+                transformer_module,
+        ) in self.trainable_transformer_modules.items():
             transformer_module.train()
             transformer_module.requires_grad_(False)
             if transformer_name in self.lora_layers:
@@ -151,32 +253,71 @@ class LoRAPipeline(ComposedPipelineBase):
         if self.lora_initialized:
             return
         self.lora_initialized = True
-        for transformer_name, transformer_module in self.trainable_transformer_modules.items(
-        ):
+        for (
+                transformer_name,
+                transformer_module,
+        ) in self.trainable_transformer_modules.items():
             converted_count = 0
+            # init bookkeeping structures
             if transformer_name not in self.lora_layers:
-                self.lora_layers[transformer_name] = {}
-            logger.info("Converting %s to LoRA Transformer", transformer_name)
-            for name, layer in transformer_module.named_modules():
-                if not self.is_target_layer(name):
-                    continue
-
-                excluded = False
-                for exclude_layer in self.exclude_lora_layers[transformer_name]:
-                    if exclude_layer in name:
-                        excluded = True
+                # get block list
+                block_list = []
+                for name, submodule in transformer_module.named_children():
+                    if isinstance(submodule, nn.ModuleList):
+                        block_list = [(f"{name}.{i}", m)
+                                      for i, m in enumerate(submodule)]
                         break
-                if excluded:
-                    continue
+                self.lora_layers[transformer_name] = LoRAModelLayers(block_list)
+            logger.info("Converting %s to LoRA Transformer", transformer_name)
+            # scan every module and convert to LoRA layer if applicable
 
-                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[transformer_name][name] = layer
-                    replace_submodule(transformer_module, name, layer)
-                    converted_count += 1
+            for block_name, block_modules in _named_module_by_prefix(
+                    transformer_module,
+                    list(self.lora_layers[transformer_name].block_mapping),
+            ):
+                if block_name is not None and (
+                        not self.fastvideo_args.training_mode
+                        and self.fastvideo_args.dit_layerwise_offload):
+                    scope_ctx = _get_hook_ctx(
+                        self.lora_layers[transformer_name].
+                        block_mapping[block_name])
+                else:
+                    scope_ctx = nullcontext()
+                with scope_ctx:
+                    for name, layer in block_modules:
+                        if not self.is_target_layer(name):
+                            continue
+
+                        excluded = False
+                        for exclude_layer in self.exclude_lora_layers[
+                                transformer_name]:
+                            if exclude_layer in name:
+                                excluded = True
+                                break
+                        if excluded:
+                            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:
+                            block_name_split = name.split(".", 2)
+                            if len(block_name_split) > 2:
+                                block_name = (block_name_split[0] + "." +
+                                              block_name_split[1])
+                            else:
+                                block_name = None
+                            if (block_name
+                                    not in self.lora_layers[transformer_name].
+                                    block_mapping):
+                                block_name = None
+                            self.lora_layers[transformer_name].add_lora_layer(
+                                block_name, name, layer)
+                            replace_submodule(transformer_module, name, layer)
+                            converted_count += 1
             logger.info("Converted %d layers to LoRA layers", converted_count)
 
     def set_lora_adapter(self,
@@ -209,7 +350,7 @@ class LoRAPipeline(ComposedPipelineBase):
 
             # Extract alpha values and weights in a single pass
             to_merge_params: defaultdict[Hashable,
-                                         dict[Any, Any]] = defaultdict(dict)
+                                         dict[Any, Any]] = (defaultdict(dict))
             for name, weight in lora_state_dict.items():
                 # Extract weights (lora_A, lora_B, and lora_alpha)
                 name = name.replace("diffusion_model.", "")
@@ -223,13 +364,14 @@ class LoRAPipeline(ComposedPipelineBase):
                     target_name, _, _ = param_names_mapping_fn(layer_name)
                     # Store alpha alongside weights with same target_name base
                     alpha_key = target_name + ".lora_alpha"
-                    self.lora_adapters[lora_nickname][alpha_key] = weight.item(
-                    ) if weight.numel() == 1 else float(weight.mean())
+                    self.lora_adapters[lora_nickname][alpha_key] = (
+                        weight.item()
+                        if weight.numel() == 1 else float(weight.mean()))
                     continue
 
                 name, _, _ = lora_param_names_mapping_fn(name)
-                target_name, merge_index, num_params_to_merge = param_names_mapping_fn(
-                    name)
+                target_name, merge_index, num_params_to_merge = (
+                    param_names_mapping_fn(name))
                 # for (in_dim, r) @ (r, out_dim), we only merge (r, out_dim * n) where n is the number of linear layers to fuse
                 # see param mapping in HunyuanVideoArchConfig
                 if merge_index is not None and "lora_B" in name:
@@ -261,45 +403,84 @@ class LoRAPipeline(ComposedPipelineBase):
 
         # Merge the new adapter
         adapted_count = 0
-        for transformer_name, transformer_lora_layers in self.lora_layers.items(
-        ):
-            for name, layer in transformer_lora_layers.items():
-                lora_A_name = name + ".lora_A"
-                lora_B_name = name + ".lora_B"
-                lora_alpha_name = name + ".lora_alpha"
-                if lora_A_name in self.lora_adapters[lora_nickname]\
-                    and lora_B_name in self.lora_adapters[lora_nickname]:
-                    # Get alpha value for this layer (defaults to None if not present)
-                    lora_A = self.lora_adapters[lora_nickname][lora_A_name]
-                    lora_B = self.lora_adapters[lora_nickname][lora_B_name]
-                    # Simple lookup - alpha stored with same naming scheme as lora_A/lora_B
-                    alpha = self.lora_adapters[lora_nickname].get(
-                        lora_alpha_name) if adapter_updated else None
-
-                    layer.set_lora_weights(
-                        lora_A,
-                        lora_B,
-                        lora_alpha=alpha,
-                        training_mode=self.fastvideo_args.training_mode,
-                        lora_path=lora_path)
-                    adapted_count += 1
-                else:
-                    if rank == 0:
-                        logger.warning(
-                            "LoRA adapter %s does not contain the weights for layer %s. LoRA will not be applied to it.",
-                            lora_path, name)
-                    layer.disable_lora = True
-        logger.info("Rank %d: LoRA adapter %s applied to %d layers", rank,
-                    lora_path, adapted_count)
+        for (
+                transformer_name,
+                transformer_lora_layers,
+        ) in self.lora_layers.items():
+            for (
+                    module,
+                    layers,
+            ) in transformer_lora_layers.lora_layers_by_block():
+                with _get_hook_ctx(module):
+                    for name, layer in layers.items():
+                        lora_A_name = name + ".lora_A"
+                        lora_B_name = name + ".lora_B"
+                        lora_alpha_name = name + ".lora_alpha"
+                        if (lora_A_name in self.lora_adapters[lora_nickname]
+                                and lora_B_name
+                                in self.lora_adapters[lora_nickname]):
+                            # Get alpha value for this layer (defaults to None if not present)
+                            lora_A = self.lora_adapters[lora_nickname][
+                                lora_A_name]
+                            lora_B = self.lora_adapters[lora_nickname][
+                                lora_B_name]
+                            # Simple lookup - alpha stored with same naming scheme as lora_A/lora_B
+                            alpha = (self.lora_adapters[lora_nickname].get(
+                                lora_alpha_name) if adapter_updated else None)
+                            try:
+                                layer.set_lora_weights(
+                                    lora_A,
+                                    lora_B,
+                                    lora_alpha=alpha,
+                                    training_mode=self.fastvideo_args.
+                                    training_mode,
+                                    lora_path=lora_path,
+                                )
+                            except Exception as e:
+                                logger.error(
+                                    "Error setting LoRA weights for layer %s: %s",
+                                    name,
+                                    str(e),
+                                )
+                                raise e
+                            adapted_count += 1
+                        else:
+                            if rank == 0:
+                                logger.warning(
+                                    "LoRA adapter %s does not contain the weights for layer %s. LoRA will not be applied to it.",
+                                    lora_path,
+                                    name,
+                                )
+                            layer.disable_lora = True
+        logger.info(
+            "Rank %d: LoRA adapter %s applied to %d layers",
+            rank,
+            lora_path,
+            adapted_count,
+        )
 
     def merge_lora_weights(self) -> None:
-        for transformer_name, transformer_lora_layers in self.lora_layers.items(
-        ):
-            for name, layer in transformer_lora_layers.items():
-                layer.merge_lora_weights()
+        for (
+                transformer_name,
+                transformer_lora_layers,
+        ) in self.lora_layers.items():
+            for (
+                    module,
+                    layers,
+            ) in transformer_lora_layers.lora_layers_by_block():
+                with _get_hook_ctx(module):
+                    for name, layer in layers.items():
+                        layer.merge_lora_weights()
 
     def unmerge_lora_weights(self) -> None:
-        for transformer_name, transformer_lora_layers in self.lora_layers.items(
-        ):
-            for name, layer in transformer_lora_layers.items():
-                layer.unmerge_lora_weights()
+        for (
+                transformer_name,
+                transformer_lora_layers,
+        ) in self.lora_layers.items():
+            for (
+                    module,
+                    layers,
+            ) in transformer_lora_layers.lora_layers_by_block():
+                with _get_hook_ctx(module):
+                    for name, layer in layers.items():
+                        layer.unmerge_lora_weights()

fastvideo/pipelines/pipeline_batch_info.py

@@ -67,6 +67,21 @@ class ForwardBatch:
     execution, allowing methods to update specific components without needing
     to manage numerous individual parameters.
     """
+
+    @dataclass
+    class RLData:
+        """RL-specific data collection options and outputs."""
+        enabled: bool = False
+        collect_log_probs: bool = True
+        collect_kl: bool = False
+        kl_reward: float = 0.0
+        store_trajectory: bool = True
+        keep_trajectory_on_cpu: bool = False
+        log_probs: torch.Tensor | None = None
+        kl: torch.Tensor | None = None
+        trajectory_latents: torch.Tensor | None = None
+        trajectory_timesteps: torch.Tensor | None = None
+
     # TODO(will): double check that args are separate from fastvideo_args
     # properly. Also maybe think about providing an abstraction for pipeline
     # specific arguments.
@@ -197,6 +212,9 @@ class ForwardBatch:
     logging_info: PipelineLoggingInfo = field(
         default_factory=PipelineLoggingInfo)
 
+    # RL data collection
+    rl_data: "ForwardBatch.RLData" = field(default_factory=RLData)
+
     def __post_init__(self):
         """Initialize dependent fields after dataclass initialization."""
 
@@ -267,6 +285,36 @@ class TrainingBatch:
     latent_vis_dict: dict[str, Any] = field(default_factory=dict)
     fake_score_latent_vis_dict: dict[str, Any] = field(default_factory=dict)
 
+    # RL/GRPO-specific attributes
+    reward_scores: torch.Tensor | None = None  # Computed rewards from reward models
+    log_probs: torch.Tensor | None = None  # Current policy log probabilities [B, num_steps] or [B]
+    old_log_probs: torch.Tensor | None = None  # Old policy log probs (for importance ratio) [B, num_steps] or [B]
+    advantages: torch.Tensor | None = None  # GAE advantages [B, num_steps] or [B]
+    returns: torch.Tensor | None = None  # TD returns (advantages + values) [B, num_steps] or [B]
+    values: torch.Tensor | None = None  # Value function predictions [B]
+    old_values: torch.Tensor | None = None  # Old value predictions (for clipping) [B]
+
+    # GRPO sampling-specific attributes
+    kl: torch.Tensor | None = None  # KL divergences from sampling [B, num_steps] (if kl_reward > 0)
+    prompt_ids: torch.Tensor | None = None  # Prompt token IDs for stat tracking [B, seq_len]
+    prompt_embeds: torch.Tensor | None = None  # Prompt embeddings used in sampling [B, seq_len, hidden_dim]
+    negative_prompt_embeds: torch.Tensor | None = None  # Negative prompt embeddings for CFG [B, seq_len, hidden_dim]
+
+    # RL loss components
+    policy_loss: float = 0.0  # GRPO/PPO policy loss
+    value_loss: float = 0.0  # Value function loss
+    kl_divergence: float = 0.0  # KL(new_policy || old_policy)
+    importance_ratio: float = 1.0  # exp(log_prob - old_log_prob)
+    clip_fraction: float = 0.0  # Fraction of ratios that were clipped
+
+    # RL metrics
+    advantage_mean: float = 0.0  # Mean advantage (should be ~0 after normalization)
+    advantage_std: float = 1.0  # Std of advantages
+    reward_mean: float = 0.0  # Mean reward across batch
+    reward_std: float = 0.0  # Std of rewards
+    value_mean: float = 0.0  # Mean value prediction
+    entropy: float = 0.0  # Policy entropy (for exploration)
+
 
 @dataclass
 class PreprocessBatch(ForwardBatch):

fastvideo/pipelines/pipeline_registry.py

@@ -24,13 +24,17 @@ _PIPELINE_NAME_TO_ARCHITECTURE_NAME: dict[str, str] = {
     "WanVideoToVideoPipeline": "wan",
     "WanCausalDMDPipeline": "wan",
     "TurboDiffusionPipeline": "turbodiffusion",
+    "TurboDiffusionI2VPipeline": "turbodiffusion",
     "StepVideoPipeline": "stepvideo",
     "HunyuanVideoPipeline": "hunyuan",
     "HunyuanVideo15Pipeline": "hunyuan15",
     "Cosmos2VideoToWorldPipeline": "cosmos",
+    "Cosmos2_5Pipeline": "cosmos",
     "MatrixGamePipeline": "matrixgame",
     "MatrixGameCausalDMDPipeline": "matrixgame",
     "LongCatPipeline": "longcat",
+    "LongCatImageToVideoPipeline": "longcat",
+    "LongCatVideoContinuationPipeline": "longcat",
 }
 
 _PREPROCESS_WORKLOAD_TYPE_TO_PIPELINE_NAME: dict[WorkloadType, str] = {

fastvideo/pipelines/stages/__init__.py

@@ -10,7 +10,8 @@ from fastvideo.pipelines.stages.base import PipelineStage
 from fastvideo.pipelines.stages.causal_denoising import CausalDMDDenosingStage
 from fastvideo.pipelines.stages.conditioning import ConditioningStage
 from fastvideo.pipelines.stages.decoding import DecodingStage
-from fastvideo.pipelines.stages.denoising import (CosmosDenoisingStage,
+from fastvideo.pipelines.stages.denoising import (Cosmos25DenoisingStage,
+                                                  CosmosDenoisingStage,
                                                   DenoisingStage,
                                                   DmdDenoisingStage)
 from fastvideo.pipelines.stages.encoding import EncodingStage
@@ -19,27 +20,37 @@ from fastvideo.pipelines.stages.image_encoding import (
     ImageVAEEncodingStage, VideoVAEEncodingStage, Hy15ImageEncodingStage)
 from fastvideo.pipelines.stages.input_validation import InputValidationStage
 from fastvideo.pipelines.stages.latent_preparation import (
-    CosmosLatentPreparationStage, LatentPreparationStage)
+    Cosmos25LatentPreparationStage, CosmosLatentPreparationStage,
+    LatentPreparationStage)
 from fastvideo.pipelines.stages.matrixgame_denoising import (
     MatrixGameCausalDenoisingStage)
 from fastvideo.pipelines.stages.stepvideo_encoding import (
     StepvideoPromptEncodingStage)
-from fastvideo.pipelines.stages.text_encoding import TextEncodingStage
+from fastvideo.pipelines.stages.text_encoding import (Cosmos25TextEncodingStage,
+                                                      TextEncodingStage)
 from fastvideo.pipelines.stages.timestep_preparation import (
-    TimestepPreparationStage)
+    Cosmos25TimestepPreparationStage, TimestepPreparationStage)
+
+# LongCat stages
+from fastvideo.pipelines.stages.longcat_video_vae_encoding import LongCatVideoVAEEncodingStage
+from fastvideo.pipelines.stages.longcat_kv_cache_init import LongCatKVCacheInitStage
+from fastvideo.pipelines.stages.longcat_vc_denoising import LongCatVCDenoisingStage
 
 __all__ = [
     "PipelineStage",
     "InputValidationStage",
     "TimestepPreparationStage",
+    "Cosmos25TimestepPreparationStage",
     "LatentPreparationStage",
     "CosmosLatentPreparationStage",
+    "Cosmos25LatentPreparationStage",
     "ConditioningStage",
     "DenoisingStage",
     "DmdDenoisingStage",
     "CausalDMDDenosingStage",
     "MatrixGameCausalDenoisingStage",
     "CosmosDenoisingStage",
+    "Cosmos25DenoisingStage",
     "EncodingStage",
     "DecodingStage",
     "ImageEncodingStage",
@@ -49,5 +60,10 @@ __all__ = [
     "ImageVAEEncodingStage",
     "VideoVAEEncodingStage",
     "TextEncodingStage",
+    "Cosmos25TextEncodingStage",
     "StepvideoPromptEncodingStage",
+    # LongCat stages
+    "LongCatVideoVAEEncodingStage",
+    "LongCatKVCacheInitStage",
+    "LongCatVCDenoisingStage",
 ]

fastvideo/pipelines/stages/decoding.py

@@ -51,39 +51,42 @@ class DecodingStage(PipelineStage):
         return result
 
     def _denormalize_latents(self, latents: torch.Tensor) -> torch.Tensor:
-        # denormalization for MatrixGame VAE
-        # z = z * std + mean during decode
-        if (hasattr(self.vae.config, 'latents_mean')
-                and hasattr(self.vae.config, 'latents_std')):
-            # Convert config values to tensors
-            latents_mean = torch.tensor(self.vae.config.latents_mean,
+        """Convert normalized latents into the VAE's expected latent space."""
+        # Some VAEs handle latent (de)normalization internally.
+        if bool(getattr(self.vae, "handles_latent_denorm", False)):
+            return latents
+
+        cfg = getattr(self.vae, "config", None)
+
+        # MatrixGame-style: z = z * std + mean
+        if (cfg is not None and hasattr(cfg, "latents_mean")
+                and hasattr(cfg, "latents_std")):
+            latents_mean = torch.tensor(cfg.latents_mean,
                                         device=latents.device,
                                         dtype=latents.dtype).view(
                                             1, -1, 1, 1, 1)
-
-            latents_std = torch.tensor(self.vae.config.latents_std,
+            latents_std = torch.tensor(cfg.latents_std,
                                        device=latents.device,
                                        dtype=latents.dtype).view(
                                            1, -1, 1, 1, 1)
+            return latents * latents_std + latents_mean
 
-            # Apply denormalization: z = z * std + mean
-            latents = latents * latents_std + latents_mean
-        elif hasattr(self.vae, 'scaling_factor'):
-            # Standard VAE scaling
+        # Diffusers-style: scaling_factor (+ optional shift_factor)
+        if hasattr(self.vae, "scaling_factor"):
             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
 
-            # Apply shifting if needed
-            if (hasattr(self.vae, "shift_factor")
-                    and self.vae.shift_factor is not None):
+            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)
+                    latents = latents + self.vae.shift_factor.to(
+                        latents.device, latents.dtype)
                 else:
-                    latents += self.vae.shift_factor
+                    latents = latents + self.vae.shift_factor
+
         return latents
 
     @torch.no_grad()
@@ -273,4 +276,4 @@ class DecodingStage(PipelineStage):
                 del pipeline.modules["vae"]
             fastvideo_args.model_loaded["vae"] = False
 
-        return batch
+        return batch

fastvideo/pipelines/stages/denoising.py

@@ -4,11 +4,14 @@ Denoising stage for diffusion pipelines.
 """
 
 import inspect
+import math
 import weakref
 from collections.abc import Iterable
+from contextlib import nullcontext
 from typing import Any
 
 import torch
+from diffusers.utils.torch_utils import randn_tensor
 from tqdm.auto import tqdm
 
 from fastvideo.attention import get_attn_backend
@@ -52,6 +55,84 @@ except ImportError:
 logger = init_logger(__name__)
 
 
+def sde_step_with_logprob(
+    scheduler,
+    model_output: torch.FloatTensor,
+    timestep: float | torch.FloatTensor,
+    sample: torch.FloatTensor,
+    prev_sample: torch.FloatTensor | None = None,
+    generator: torch.Generator | None = None,
+    deterministic: bool = False,
+    return_pixel_log_prob: bool = False,
+    return_dt_and_std_dev_t: bool = False
+) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor, ...]:
+    """
+    Predict the sample from the previous timestep by reversing the SDE and
+    compute log probabilities for the transition.
+    """
+    if isinstance(timestep, torch.Tensor):
+        if timestep.ndim == 0:
+            timestep = timestep.unsqueeze(0)
+        step_indices = [
+            scheduler.index_for_timestep(t.item()) for t in timestep
+        ]
+    else:
+        step_indices = [scheduler.index_for_timestep(timestep)]
+
+    prev_step_indices = [step + 1 for step in step_indices]
+
+    sigmas = scheduler.sigmas.to(sample.device, sample.dtype)
+    sigma = sigmas[step_indices].view(-1, 1, 1, 1, 1)
+    sigma_prev = sigmas[prev_step_indices].view(-1, 1, 1, 1, 1)
+    sigma_max = sigmas[0].item()
+    sigma_min = sigmas[-1].item()
+
+    dt = sigma_prev - sigma
+
+    std_dev_t = sigma_min + (sigma_max - sigma_min) * sigma
+    prev_sample_mean = (sample * (1 + std_dev_t**2 / (2 * sigma) * dt) +
+                        model_output * (1 + std_dev_t**2 * (1 - sigma) /
+                                        (2 * sigma)) * dt)
+
+    if prev_sample is not None and generator is not None:
+        raise ValueError(
+            "Cannot pass both generator and prev_sample. Please make sure that either `generator` or"
+            " `prev_sample` stays `None`.")
+
+    if prev_sample is None:
+        variance_noise = randn_tensor(
+            model_output.shape,
+            generator=generator,
+            device=model_output.device,
+            dtype=model_output.dtype,
+        )
+        sqrt_dt = torch.sqrt(-1 * dt)
+        prev_sample = prev_sample_mean + std_dev_t * sqrt_dt * variance_noise
+    else:
+        sqrt_dt = torch.sqrt(-1 * dt)
+
+    if deterministic:
+        prev_sample = sample + dt * model_output
+        sqrt_dt = torch.sqrt(-1 * dt)
+
+    if return_pixel_log_prob:
+        raise NotImplementedError(
+            "Pixel-level log prob is not supported in this helper.")
+
+    std_dev_sqrt_dt = std_dev_t * sqrt_dt
+    log_prob = (
+        -((prev_sample.detach() - prev_sample_mean)**2) /
+        (2 *
+         (std_dev_sqrt_dt**2)) - torch.log(std_dev_sqrt_dt + 1e-8) - torch.log(
+             torch.sqrt(2 * torch.as_tensor(math.pi, device=sample.device))))
+
+    log_prob = log_prob.mean(dim=tuple(range(1, log_prob.ndim)))
+
+    if return_dt_and_std_dev_t:
+        return prev_sample, log_prob, prev_sample_mean, std_dev_t, sqrt_dt
+    return prev_sample, log_prob, prev_sample_mean, std_dev_t * sqrt_dt
+
+
 class DenoisingStage(PipelineStage):
     """
     Stage for running the denoising loop in diffusion pipelines.
@@ -203,9 +284,11 @@ class DenoisingStage(PipelineStage):
         else:
             boundary_timestep = None
         latent_model_input = latents.to(target_dtype)
-        assert latent_model_input.shape[0] == 1, "only support batch size 1"
+        rl_data = batch.rl_data if batch.rl_data and batch.rl_data.enabled else None
 
         if fastvideo_args.pipeline_config.ti2v_task and batch.pil_image is not None:
+            assert latent_model_input.shape[
+                0] == 1, "TI2V task only supports batch size 1"
             # TI2V directly replaces the first frame of the latent with
             # the image latent instead of appending along the channel dim
             assert batch.image_latent is None, "TI2V task should not have image latents"
@@ -243,6 +326,12 @@ class DenoisingStage(PipelineStage):
         # Initialize lists for ODE trajectory
         trajectory_timesteps: list[torch.Tensor] = []
         trajectory_latents: list[torch.Tensor] = []
+        rl_timesteps: list[torch.Tensor] = []
+        rl_latents: list[torch.Tensor] = []
+        rl_log_probs: list[torch.Tensor] = []
+        rl_kl: list[torch.Tensor] = []
+        if rl_data is not None and rl_data.store_trajectory:
+            rl_latents.append(latents)
 
         # Run denoising loop
         with self.progress_bar(total=num_inference_steps) as progress_bar:
@@ -329,6 +418,24 @@ class DenoisingStage(PipelineStage):
                     1000.0 if fastvideo_args.pipeline_config.embedded_cfg_scale
                     is not None else None)
 
+                def run_transformer(model, encoder_hidden_states, cond_kwargs,
+                                    is_cfg_negative: bool):
+                    batch.is_cfg_negative = is_cfg_negative
+                    with set_forward_context(
+                            current_timestep=i,
+                            attn_metadata=attn_metadata,
+                            forward_batch=batch,
+                    ):
+                        return model(
+                            latent_model_input,
+                            encoder_hidden_states,
+                            t_expand,
+                            guidance=guidance_expand,
+                            **image_kwargs,
+                            **cond_kwargs,
+                            **action_kwargs,
+                        )
+
                 # Predict noise residual
                 with torch.autocast(device_type="cuda",
                                     dtype=target_dtype,
@@ -390,40 +497,13 @@ class DenoisingStage(PipelineStage):
                     # support torch dynamo compilation. They pass in
                     # attn_metadata, vllm_config, and num_tokens. We can pass in
                     # fastvideo_args or training_args, and attn_metadata.
-                    batch.is_cfg_negative = False
-                    with set_forward_context(
-                            current_timestep=i,
-                            attn_metadata=attn_metadata,
-                            forward_batch=batch,
-                            # fastvideo_args=fastvideo_args
-                    ):
-                        # Run transformer
-                        noise_pred = current_model(
-                            latent_model_input,
-                            prompt_embeds,
-                            t_expand,
-                            guidance=guidance_expand,
-                            **image_kwargs,
-                            **pos_cond_kwargs,
-                            **action_kwargs,
-                        )
+                    noise_pred = run_transformer(current_model, prompt_embeds,
+                                                 pos_cond_kwargs, False)
 
                     if batch.do_classifier_free_guidance:
-                        batch.is_cfg_negative = True
-                        with set_forward_context(
-                                current_timestep=i,
-                                attn_metadata=attn_metadata,
-                                forward_batch=batch,
-                        ):
-                            noise_pred_uncond = current_model(
-                                latent_model_input,
-                                neg_prompt_embeds,
-                                t_expand,
-                                guidance=guidance_expand,
-                                **image_kwargs,
-                                **neg_cond_kwargs,
-                                **action_kwargs,
-                            )
+                        noise_pred_uncond = run_transformer(
+                            current_model, neg_prompt_embeds, neg_cond_kwargs,
+                            True)
 
                         noise_pred_text = noise_pred
                         noise_pred = noise_pred_uncond + current_guidance_scale * (
@@ -438,11 +518,58 @@ class DenoisingStage(PipelineStage):
                                 guidance_rescale=batch.guidance_rescale,
                             )
                     # Compute the previous noisy sample
+                    prev_latents = latents
                     latents = self.scheduler.step(noise_pred,
                                                   t,
                                                   latents,
                                                   **extra_step_kwargs,
                                                   return_dict=False)[0]
+                    if rl_data is not None:
+                        if rl_data.collect_log_probs:
+                            _, log_prob, prev_latents_mean, std_dev_t, _ = sde_step_with_logprob(
+                                self.scheduler,
+                                noise_pred.float(),
+                                t,
+                                prev_latents.float(),
+                                prev_sample=latents.float(),
+                                deterministic=False,
+                                return_dt_and_std_dev_t=True,
+                            )
+                            rl_log_probs.append(log_prob)
+
+                        if rl_data.collect_kl and rl_data.kl_reward > 0:
+                            adapter_ctx = nullcontext()
+                            if hasattr(current_model, "disable_adapter"):
+                                adapter_ctx = current_model.disable_adapter()
+                            with adapter_ctx:
+                                noise_pred_ref = run_transformer(
+                                    current_model, prompt_embeds,
+                                    pos_cond_kwargs, False)
+                                if batch.do_classifier_free_guidance:
+                                    noise_pred_uncond_ref = run_transformer(
+                                        current_model, neg_prompt_embeds,
+                                        neg_cond_kwargs, True)
+                                    noise_pred_text_ref = noise_pred_ref
+                                    noise_pred_ref = noise_pred_uncond_ref + current_guidance_scale * (
+                                        noise_pred_text_ref -
+                                        noise_pred_uncond_ref)
+                            _, _, prev_latents_mean_ref, std_dev_t_ref, _ = sde_step_with_logprob(
+                                self.scheduler,
+                                noise_pred_ref.float(),
+                                t,
+                                prev_latents.float(),
+                                prev_sample=latents.float(),
+                                deterministic=False,
+                                return_dt_and_std_dev_t=True,
+                            )
+                            if not torch.allclose(std_dev_t, std_dev_t_ref):
+                                logger.warning(
+                                    "std_dev_t mismatch in RL KL computation at step %s",
+                                    i)
+                            kl = (prev_latents_mean -
+                                  prev_latents_mean_ref)**2 / (2 * std_dev_t**2)
+                            kl = kl.mean(dim=tuple(range(1, kl.ndim)))
+                            rl_kl.append(kl)
                     if fastvideo_args.pipeline_config.ti2v_task and batch.pil_image is not None:
                         latents = latents.squeeze(0)
                         latents = (1. - mask2[0]) * z + mask2[0] * latents
@@ -452,6 +579,15 @@ class DenoisingStage(PipelineStage):
                 if batch.return_trajectory_latents:
                     trajectory_timesteps.append(t)
                     trajectory_latents.append(latents)
+                if rl_data is not None:
+                    rl_timesteps.append(t)
+                    if rl_data.store_trajectory:
+                        rl_latents.append(latents)
+                    if rl_data.collect_kl and rl_data.kl_reward <= 0:
+                        rl_kl.append(
+                            torch.zeros(latents.shape[0],
+                                        device=latents.device,
+                                        dtype=latents.dtype))
 
                 # Update progress bar
                 if i == len(timesteps) - 1 or (
@@ -472,6 +608,25 @@ class DenoisingStage(PipelineStage):
         if trajectory_tensor is not None and trajectory_timesteps_tensor is not None:
             batch.trajectory_timesteps = trajectory_timesteps_tensor.cpu()
             batch.trajectory_latents = trajectory_tensor.cpu()
+        if rl_data is not None:
+            if rl_timesteps:
+                rl_data.trajectory_timesteps = torch.stack(rl_timesteps, dim=0)
+                if rl_data.keep_trajectory_on_cpu:
+                    rl_data.trajectory_timesteps = rl_data.trajectory_timesteps.cpu(
+                    )
+            if rl_data.store_trajectory and rl_latents:
+                rl_data.trajectory_latents = torch.stack(rl_latents, dim=1)
+                if rl_data.keep_trajectory_on_cpu:
+                    rl_data.trajectory_latents = rl_data.trajectory_latents.cpu(
+                    )
+            if rl_log_probs:
+                rl_data.log_probs = torch.stack(rl_log_probs, dim=1)
+                if rl_data.keep_trajectory_on_cpu:
+                    rl_data.log_probs = rl_data.log_probs.cpu()
+            if rl_kl:
+                rl_data.kl = torch.stack(rl_kl, dim=1)
+                if rl_data.keep_trajectory_on_cpu:
+                    rl_data.kl = rl_data.kl.cpu()
 
         # Update batch with final latents
         batch.latents = latents
@@ -1018,6 +1173,145 @@ class CosmosDenoisingStage(DenoisingStage):
         return result
 
 
+class Cosmos25DenoisingStage(CosmosDenoisingStage):
+    """Denoising stage for Cosmos 2.5 DiT (expects 1D/2D timestep, not 5D)."""
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        pipeline = self.pipeline() if self.pipeline else None
+        if not fastvideo_args.model_loaded["transformer"]:
+            loader = TransformerLoader()
+            self.transformer = loader.load(
+                fastvideo_args.model_paths["transformer"], fastvideo_args)
+            if pipeline:
+                pipeline.add_module("transformer", self.transformer)
+            fastvideo_args.model_loaded["transformer"] = True
+
+        extra_step_kwargs = self.prepare_extra_func_kwargs(
+            self.scheduler.step,
+            {
+                "generator": batch.generator,
+                "eta": batch.eta
+            },
+        )
+
+        if hasattr(self.transformer, 'module'):
+            transformer_dtype = next(self.transformer.module.parameters()).dtype
+        else:
+            transformer_dtype = next(self.transformer.parameters()).dtype
+        target_dtype = transformer_dtype
+        autocast_enabled = (target_dtype != torch.float32
+                            ) and not fastvideo_args.disable_autocast
+
+        latents = batch.latents
+        if latents is None:
+            raise ValueError(
+                "latents must be provided for Cosmos25DenoisingStage")
+        guidance_scale = batch.guidance_scale
+
+        # Use timesteps prepared by Cosmos25TimestepPreparationStage when available.
+        if batch.timesteps is None:
+            self.scheduler.set_timesteps(batch.num_inference_steps,
+                                         device=latents.device)
+            timesteps = self.scheduler.timesteps
+        else:
+            timesteps = batch.timesteps.to(latents.device)
+
+        # Match official behavior: pass fps as a tensor.
+        fps_val = batch.fps if isinstance(batch.fps, int | float) else 24
+        fps_tensor = torch.tensor([fps_val],
+                                  device=latents.device,
+                                  dtype=target_dtype)
+
+        # Cosmos2.5 denoises a 4D latent (C,T,H,W) and the scheduler.step expects (B,C,T,H,W).
+        latents_4d = latents[0]
+
+        # Masks from latent prep stage
+        condition_mask = batch.cond_mask.to(target_dtype) if hasattr(
+            batch, 'cond_mask') else None
+        padding_mask = batch.padding_mask.to(target_dtype) if hasattr(
+            batch, 'padding_mask') else None
+        if condition_mask is None:
+            _, t, h, w = latents_4d.shape
+            condition_mask = torch.zeros(1,
+                                         1,
+                                         t,
+                                         h,
+                                         w,
+                                         device=latents.device,
+                                         dtype=target_dtype)
+        if padding_mask is None:
+            _, _, h, w = latents_4d.shape
+            padding_mask = torch.ones(1,
+                                      1,
+                                      h,
+                                      w,
+                                      device=latents.device,
+                                      dtype=target_dtype)
+
+        # Cosmos2.5 timestep scaling (see compare_pipelines.py): t * 0.001
+        timestep_scale = 0.001
+
+        with self.progress_bar(total=len(timesteps)) as progress_bar:
+            for i, t in enumerate(timesteps):
+                t_val = float(t)
+                timestep_val = t_val * timestep_scale
+                timestep = torch.tensor([[timestep_val]],
+                                        device=latents.device,
+                                        dtype=target_dtype)
+
+                model_hidden_states = latents_4d.unsqueeze(0)
+
+                with (
+                        set_forward_context(current_timestep=int(t_val),
+                                            attn_metadata=None,
+                                            forward_batch=batch),
+                        torch.autocast(device_type="cuda",
+                                       dtype=target_dtype,
+                                       enabled=autocast_enabled),
+                ):
+                    cond_v = self.transformer(
+                        hidden_states=model_hidden_states.to(target_dtype),
+                        encoder_hidden_states=batch.prompt_embeds[0].to(
+                            target_dtype),
+                        timestep=timestep,
+                        fps=fps_tensor,
+                        condition_mask=condition_mask,
+                        padding_mask=padding_mask,
+                        return_dict=False,
+                    )[0]
+
+                    if batch.do_classifier_free_guidance and batch.negative_prompt_embeds:
+                        uncond_v = self.transformer(
+                            hidden_states=model_hidden_states.to(target_dtype),
+                            encoder_hidden_states=batch.
+                            negative_prompt_embeds[0].to(target_dtype),
+                            timestep=timestep,
+                            fps=fps_tensor,
+                            condition_mask=condition_mask,
+                            padding_mask=padding_mask,
+                            return_dict=False,
+                        )[0]
+                        v = uncond_v + guidance_scale * (cond_v - uncond_v)
+                    else:
+                        v = cond_v
+
+                prev = self.scheduler.step(v.unsqueeze(0),
+                                           t,
+                                           latents_4d.unsqueeze(0),
+                                           **extra_step_kwargs,
+                                           return_dict=False)[0]
+                latents_4d = prev.squeeze(0)
+
+                progress_bar.update()
+
+        batch.latents = latents_4d.unsqueeze(0)
+        return batch
+
+
 class DmdDenoisingStage(DenoisingStage):
     """
     Denoising stage for DMD.

fastvideo/pipelines/stages/latent_preparation.py

@@ -5,6 +5,7 @@ Latent preparation stage for diffusion pipelines.
 
 from typing import Any
 
+import numpy as np
 import torch
 from diffusers.utils.torch_utils import randn_tensor
 
@@ -427,6 +428,210 @@ class CosmosLatentPreparationStage(PipelineStage):
 
         return batch
 
+
+class Cosmos25LatentPreparationStage(CosmosLatentPreparationStage):
+    """Latent preparation for Cosmos 2.5 DiT input conventions."""
+
+    @staticmethod
+    def _arch_invariant_randn(
+        shape: tuple[int, ...],
+        *,
+        seed: int,
+        device: torch.device,
+        dtype: torch.dtype,
+    ) -> torch.Tensor:
+        """Architecture-invariant RNG (matches cosmos_predict2.misc.arch_invariant_rand)."""
+        rng = np.random.RandomState(seed)
+        arr = rng.standard_normal(shape).astype(np.float32)
+        return torch.from_numpy(arr).to(device=device, dtype=dtype)
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        # Differences vs `CosmosLatentPreparationStage`: channel convention, seed usage,
+        # and `padding_mask` for concat_padding_mask=True.
+
+        # Determine batch size
+        if isinstance(batch.prompt, list):
+            batch_size = len(batch.prompt)
+        elif batch.prompt is not None:
+            batch_size = 1
+        else:
+            batch_size = batch.prompt_embeds[0].shape[0]
+
+        batch_size *= batch.num_videos_per_prompt
+
+        # Match `compare_pipelines.py`: initialize noise in fp32, then run the
+        # denoising computation in bf16.
+        dtype = torch.float32
+        device = get_local_torch_device()
+        generator = batch.generator
+        latents = batch.latents
+        num_frames = batch.num_frames
+        height = batch.height
+        width = batch.width
+
+        if height is None or width is None:
+            raise ValueError("Height and width must be provided")
+
+        vae_scale_factor_spatial = 8
+        vae_scale_factor_temporal = 4
+
+        latent_height = height // 8
+        latent_width = width // vae_scale_factor_spatial
+        num_latent_frames = (num_frames - 1) // vae_scale_factor_temporal + 1
+
+        # Cosmos 2.5 convention: transformer in_channels == latent channels
+        num_channels_latents = self.transformer.config.in_channels
+
+        shape = (batch_size, num_channels_latents, num_latent_frames,
+                 latent_height, latent_width)
+
+        init_latents = None
+        conditioning_latents = None
+        video = None
+
+        if hasattr(batch, 'video') and batch.video is not None:
+            video = batch.video
+        elif hasattr(batch, 'pil_image') and batch.pil_image is not None:
+            vae_scale_factor_spatial = 8
+            image_processor = ImageProcessor(
+                vae_scale_factor=vae_scale_factor_spatial)
+            processed_image = image_processor.preprocess(
+                batch.pil_image, height, width)
+            video = processed_image.unsqueeze(2)
+            video = video.to(device=device, dtype=torch.bfloat16)
+        elif hasattr(
+                batch,
+                'preprocessed_image') and batch.preprocessed_image is not None:
+            if isinstance(batch.preprocessed_image, torch.Tensor):
+                if batch.preprocessed_image.dim() == 4:
+                    video = batch.preprocessed_image.unsqueeze(2)
+                elif batch.preprocessed_image.dim() == 5:
+                    video = batch.preprocessed_image
+        else:
+            logger.info(
+                "CosmosLatentPreparationStage - No video input sources found")
+
+        if video is not None:
+            num_cond_frames = video.size(2)
+            if num_cond_frames >= num_frames:
+                num_cond_latent_frames = (num_frames -
+                                          1) // vae_scale_factor_temporal + 1
+                video = video[:, :, -num_frames:]
+            else:
+                num_cond_latent_frames = (num_cond_frames -
+                                          1) // vae_scale_factor_temporal + 1
+                num_padding_frames = num_frames - num_cond_frames
+                last_frame = video[:, :, -1:]
+                padding = last_frame.repeat(1, 1, num_padding_frames, 1, 1)
+                video = torch.cat([video, padding], dim=2)
+
+            if self.vae is not None:
+                self.vae = self.vae.to(device)
+                self.vae = self.vae.to(dtype=video.dtype)
+
+                def retrieve_latents(
+                        encoder_output: Any,
+                        generator: Any | None = None) -> torch.Tensor:
+                    if hasattr(encoder_output, "latent_dist"):
+                        return encoder_output.latent_dist.sample(generator)
+                    elif hasattr(encoder_output, "latents"):
+                        return encoder_output.latents
+                    elif hasattr(encoder_output, "sample"):
+                        return encoder_output.sample(generator)
+                    elif isinstance(encoder_output, torch.Tensor):
+                        return encoder_output
+                    else:
+                        attrs = [
+                            attr for attr in dir(encoder_output)
+                            if not attr.startswith('_')
+                        ]
+                        raise AttributeError(
+                            f"Could not access latents of provided encoder_output. Available attributes: {attrs}"
+                        )
+
+                if isinstance(generator, list):
+                    init_latents = [
+                        retrieve_latents(self.vae.encode(video[i].unsqueeze(0)),
+                                         generator=torch.Generator(
+                                             device="cpu").manual_seed(100))
+                        for i in range(batch_size)
+                    ]
+                else:
+                    init_latents = [
+                        retrieve_latents(
+                            self.vae.encode(vid.unsqueeze(0)),
+                            torch.Generator(device="cpu").manual_seed(100))
+                        for vid in video
+                    ]
+
+                init_latents = torch.cat(init_latents, dim=0).to(dtype)
+
+                cfg = getattr(self.vae, "config", None)
+                if (not bool(getattr(self.vae, "handles_latent_norm", False))
+                        and cfg is not None and hasattr(cfg, 'latents_mean')
+                        and hasattr(cfg, 'latents_std')):
+                    latents_mean = torch.tensor(cfg.latents_mean).view(
+                        1, cfg.z_dim, 1, 1, 1).to(device, dtype)
+                    latents_std = torch.tensor(cfg.latents_std).view(
+                        1, cfg.z_dim, 1, 1, 1).to(device, dtype)
+                    init_latents = (init_latents - latents_mean
+                                    ) / latents_std * self.scheduler.sigma_data
+
+                conditioning_latents = init_latents
+                self.vae.to("cpu")
+        else:
+            num_cond_latent_frames = 0
+
+        if latents is None:
+            seed = int(batch.seed if batch.seed is not None else 0)
+            # Use arch-invariant RNG to match Cosmos2.5 reference sampling.
+            latents_fp32 = self._arch_invariant_randn(shape,
+                                                      seed=seed,
+                                                      device=device,
+                                                      dtype=torch.float32)
+            latents = latents_fp32.to(torch.bfloat16)
+        else:
+            # If latents are supplied, keep compute dtype consistent with Cosmos sampling.
+            latents = latents.to(device=device, dtype=torch.bfloat16)
+
+        # Cosmos2.5 starts from unit Gaussian noise (no extra sigma_max scaling).
+
+        padding_shape = (batch_size, 1, num_latent_frames, latent_height,
+                         latent_width)
+        ones_padding = latents.new_ones(padding_shape)
+        zeros_padding = latents.new_zeros(padding_shape)
+
+        cond_indicator = latents.new_zeros(1, 1, latents.size(2), 1, 1)
+        cond_indicator[:, :, :num_cond_latent_frames] = 1.0
+        cond_mask = cond_indicator * ones_padding + (
+            1 - cond_indicator) * zeros_padding
+
+        uncond_indicator = None
+        uncond_mask = None
+        if batch.do_classifier_free_guidance:
+            uncond_indicator = latents.new_zeros(1, 1, latents.size(2), 1, 1)
+            uncond_indicator[:, :, :num_cond_latent_frames] = 1.0
+            uncond_mask = uncond_indicator * ones_padding + (
+                1 - uncond_indicator) * zeros_padding
+
+        # Cosmos 2.5 requires a spatial padding mask when concat_padding_mask=True
+        padding_mask = latents.new_ones(batch_size, 1, latent_height,
+                                        latent_width)
+
+        batch.latents = latents
+        batch.raw_latent_shape = latents.shape
+        batch.conditioning_latents = conditioning_latents
+        batch.cond_indicator = cond_indicator
+        batch.uncond_indicator = uncond_indicator
+        batch.cond_mask = cond_mask
+        batch.uncond_mask = uncond_mask
+        batch.padding_mask = padding_mask
+        return batch
+
     def adjust_video_length(self, batch: ForwardBatch,
                             fastvideo_args: FastVideoArgs) -> int:
         """

fastvideo/pipelines/stages/longcat_denoising.py

@@ -64,7 +64,7 @@ class LongCatDenoisingStage(DenoisingStage):
             The batch with denoised latents.
         """
         if not fastvideo_args.model_loaded["transformer"]:
-            from fastvideo.models.model_loader import TransformerLoader
+            from fastvideo.models.loader.component_loader import TransformerLoader
             loader = TransformerLoader()
             self.transformer = loader.load(
                 fastvideo_args.model_paths["transformer"], fastvideo_args)

fastvideo/pipelines/stages/longcat_i2v_denoising.py

@@ -0,0 +1,171 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat I2V Denoising Stage with conditioning support.
+
+This stage implements Tier 3 I2V denoising:
+1. Per-frame timestep masking (timestep[:, :num_cond_latents] = 0)
+2. Passes num_cond_latents to transformer (for RoPE skipping)
+3. Selective denoising (only updates non-conditioned frames)
+4. CFG-zero optimized guidance
+"""
+
+import torch
+from tqdm import tqdm
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.forward_context import set_forward_context
+from fastvideo.logger import init_logger
+from fastvideo.models.loader.component_loader import TransformerLoader
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.stages.longcat_denoising import LongCatDenoisingStage
+
+logger = init_logger(__name__)
+
+
+class LongCatI2VDenoisingStage(LongCatDenoisingStage):
+    """
+    LongCat denoising with I2V conditioning support.
+    
+    Key modifications from base LongCat denoising:
+    1. Sets timestep=0 for conditioning frames
+    2. Passes num_cond_latents to transformer
+    3. Only applies scheduler step to non-conditioned frames
+    """
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        """Run denoising loop with I2V conditioning."""
+
+        # Load transformer if needed
+        if not fastvideo_args.model_loaded["transformer"]:
+            loader = TransformerLoader()
+            self.transformer = loader.load(
+                fastvideo_args.model_paths["transformer"], fastvideo_args)
+            fastvideo_args.model_loaded["transformer"] = True
+
+        # Setup
+        target_dtype = torch.bfloat16
+        autocast_enabled = (target_dtype != torch.float32
+                            ) and not fastvideo_args.disable_autocast
+
+        latents = batch.latents
+        timesteps = batch.timesteps
+        prompt_embeds = batch.prompt_embeds[0]
+        prompt_attention_mask = (batch.prompt_attention_mask[0]
+                                 if batch.prompt_attention_mask else None)
+        guidance_scale = batch.guidance_scale
+        do_classifier_free_guidance = batch.do_classifier_free_guidance
+
+        # Get num_cond_latents from batch
+        num_cond_latents = getattr(batch, 'num_cond_latents', 0)
+
+        if num_cond_latents > 0:
+            logger.info("I2V Denoising: num_cond_latents=%s, latent_shape=%s",
+                        num_cond_latents, latents.shape)
+
+        # Prepare negative prompts for CFG
+        if do_classifier_free_guidance:
+            negative_prompt_embeds = batch.negative_prompt_embeds[0]
+            negative_prompt_attention_mask = (batch.negative_attention_mask[0]
+                                              if batch.negative_attention_mask
+                                              else None)
+
+            prompt_embeds_combined = torch.cat(
+                [negative_prompt_embeds, prompt_embeds], dim=0)
+            if prompt_attention_mask is not None:
+                prompt_attention_mask_combined = torch.cat(
+                    [negative_prompt_attention_mask, prompt_attention_mask],
+                    dim=0)
+            else:
+                prompt_attention_mask_combined = None
+        else:
+            prompt_embeds_combined = prompt_embeds
+            prompt_attention_mask_combined = prompt_attention_mask
+
+        # Denoising loop
+        num_inference_steps = len(timesteps)
+
+        with tqdm(total=num_inference_steps,
+                  desc="I2V Denoising") as progress_bar:
+            for i, t in enumerate(timesteps):
+
+                # 1. Expand latents for CFG
+                if do_classifier_free_guidance:
+                    latent_model_input = torch.cat([latents] * 2)
+                else:
+                    latent_model_input = latents
+
+                latent_model_input = latent_model_input.to(target_dtype)
+
+                # 2. Expand timestep to match batch size
+                timestep = t.expand(
+                    latent_model_input.shape[0]).to(target_dtype)
+
+                # 3. CRITICAL: Expand timestep to temporal dimension
+                # and set conditioning frames to timestep=0
+                timestep = timestep.unsqueeze(-1).repeat(
+                    1, latent_model_input.shape[2])
+
+                # Mark conditioning frames as clean (timestep=0)
+                if num_cond_latents > 0:
+                    timestep[:, :num_cond_latents] = 0
+
+                # 4. Run transformer with num_cond_latents
+                batch.is_cfg_negative = False
+                with set_forward_context(
+                        current_timestep=i,
+                        attn_metadata=None,
+                        forward_batch=batch,
+                ), torch.autocast(device_type='cuda',
+                                  dtype=target_dtype,
+                                  enabled=autocast_enabled):
+                    noise_pred = self.transformer(
+                        hidden_states=latent_model_input,
+                        encoder_hidden_states=prompt_embeds_combined,
+                        timestep=timestep,
+                        encoder_attention_mask=prompt_attention_mask_combined,
+                        num_cond_latents=num_cond_latents,
+                    )
+
+                # 5. Apply CFG with optimized scale
+                if do_classifier_free_guidance:
+                    noise_pred_uncond, noise_pred_cond = noise_pred.chunk(2)
+
+                    B = noise_pred_cond.shape[0]
+                    positive = noise_pred_cond.reshape(B, -1)
+                    negative = noise_pred_uncond.reshape(B, -1)
+
+                    # CFG-zero optimized scale
+                    st_star = self.optimized_scale(positive, negative)
+                    st_star = st_star.view(B, 1, 1, 1, 1)
+
+                    noise_pred = (
+                        noise_pred_uncond * st_star + guidance_scale *
+                        (noise_pred_cond - noise_pred_uncond * st_star))
+
+                # 6. CRITICAL: Negate for flow matching scheduler
+                noise_pred = -noise_pred
+
+                # 7. CRITICAL: Only update non-conditioned frames
+                # The conditioning frames stay FIXED throughout denoising
+                if num_cond_latents > 0:
+                    latents[:, :, num_cond_latents:] = self.scheduler.step(
+                        noise_pred[:, :, num_cond_latents:],
+                        t,
+                        latents[:, :, num_cond_latents:],
+                        return_dict=False)[0]
+                else:
+                    # No conditioning, update all frames
+                    latents = self.scheduler.step(noise_pred,
+                                                  t,
+                                                  latents,
+                                                  return_dict=False)[0]
+
+                progress_bar.update()
+
+        # Update batch with denoised latents
+        batch.latents = latents
+        return batch

fastvideo/pipelines/stages/longcat_i2v_latent_preparation.py

@@ -0,0 +1,105 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat I2V Latent Preparation Stage.
+
+This stage prepares latents with image conditioning for the first frame.
+"""
+
+import torch
+
+from fastvideo.distributed import get_local_torch_device
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.stages.latent_preparation import LatentPreparationStage
+
+logger = init_logger(__name__)
+
+
+class LongCatI2VLatentPreparationStage(LatentPreparationStage):
+    """
+    Prepare latents with image conditioning for first frame.
+    
+    This stage:
+    1. Generates random noise for all frames
+    2. Replaces first latent frame with encoded image latent
+    3. Marks conditioning information in batch
+    """
+
+    # Uses parent __init__ - no need for additional constructor
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        """Prepare latents with I2V conditioning."""
+
+        # IMPORTANT: Skip if latents already prepared (e.g., by refinement init stage)
+        # The refine_init stage encodes stage1 video and mixes with noise - don't overwrite!
+        if batch.latents is not None:
+            logger.info(
+                "I2V Latent Prep: Skipping - latents already prepared "
+                "(shape=%s), likely from refinement stage", batch.latents.shape)
+            return batch
+
+        # 1. Calculate dimensions
+        num_frames = batch.num_frames
+        height = batch.height
+        width = batch.width
+
+        # Get VAE compression factors
+        # IMPORTANT: Use VAE's temporal compression (4), NOT transformer's patch_size[0] (1)
+        vae_temporal_scale = fastvideo_args.pipeline_config.vae_config.arch_config.scale_factor_temporal
+        vae_spatial_scale = fastvideo_args.pipeline_config.vae_config.arch_config.scale_factor_spatial
+
+        num_latent_frames = (num_frames - 1) // vae_temporal_scale + 1
+        latent_height = height // vae_spatial_scale
+        latent_width = width // vae_spatial_scale
+
+        num_channels = self.transformer.config.in_channels
+
+        logger.info(
+            "I2V Latent Prep: num_frames=%s, num_latent_frames=%s "
+            "(vae_temporal_scale=%s), latent_shape=(%s, %s)", num_frames,
+            num_latent_frames, vae_temporal_scale, latent_height, latent_width)
+
+        # 2. Generate random noise for all frames
+        # batch_size might not be set, default to 1
+        batch_size = batch.batch_size if batch.batch_size is not None else 1
+        shape = (batch_size, num_channels, num_latent_frames, latent_height,
+                 latent_width)
+
+        # Handle generator - may be a list for batch handling
+        generator = batch.generator
+        if isinstance(generator, list):
+            generator = generator[0] if generator else None
+
+        # torch.randn requires specific argument order: size, generator, dtype
+        latents = torch.randn(*shape,
+                              generator=generator).to(get_local_torch_device(),
+                                                      dtype=torch.float32)
+
+        # 3. Replace first frame with conditioned image latent
+        if batch.image_latent is not None:
+            num_cond_latents = batch.num_cond_latents
+            latents[:, :, :
+                    num_cond_latents] = batch.image_latent[:, :, :
+                                                           num_cond_latents]
+
+            logger.info(
+                "I2V: Replaced first %s latent frame(s) with image conditioning",
+                num_cond_latents)
+        else:
+            logger.warning(
+                "No image_latent found in batch, proceeding without conditioning"
+            )
+
+        # 4. Store in batch
+        batch.latents = latents
+
+        # Required by base class output validator
+        batch.raw_latent_shape = (num_latent_frames, latent_height,
+                                  latent_width)
+
+        return batch

fastvideo/pipelines/stages/longcat_image_vae_encoding.py

@@ -0,0 +1,162 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat Image VAE Encoding Stage for I2V generation.
+
+This stage handles encoding a single input image to latent space with
+LongCat-specific normalization for I2V conditioning.
+"""
+
+import PIL
+import torch
+
+from fastvideo.distributed import get_local_torch_device
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.vision_utils import (normalize, numpy_to_pt, pil_to_numpy,
+                                           resize)
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.stages.base import PipelineStage
+from fastvideo.utils import PRECISION_TO_TYPE
+
+logger = init_logger(__name__)
+
+
+class LongCatImageVAEEncodingStage(PipelineStage):
+    """
+    Encode input image to latent space for I2V conditioning.
+    
+    This stage:
+    1. Preprocesses image to match target dimensions
+    2. Encodes via VAE to latent space
+    3. Applies LongCat-specific normalization
+    4. Stores latent and calculates num_cond_latents
+    """
+
+    def __init__(self, vae):
+        super().__init__()
+        self.vae = vae
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        """Encode image to latent for I2V conditioning."""
+
+        # Skip image encoding for refinement tasks - we're refining an existing video
+        if getattr(batch, 'stage1_video', None) is not None or getattr(
+                batch, 'refine_from', None) is not None:
+            logger.info(
+                "Skipping image encoding - refinement mode (using stage1_video)"
+            )
+            return batch
+
+        # 1. Get image from batch
+        image = batch.pil_image  # PIL.Image
+        if image is None:
+            raise ValueError("pil_image must be provided for I2V")
+
+        if not isinstance(image, PIL.Image.Image):
+            raise TypeError(f"pil_image must be PIL.Image, got {type(image)}")
+
+        # 2. Get target dimensions
+        height = batch.height
+        width = batch.width
+
+        if height is None or width is None:
+            raise ValueError("height and width must be set for I2V")
+
+        # 3. Preprocess image
+        image = resize(image, height, width, resize_mode="default")
+        image = pil_to_numpy(image)
+        image = numpy_to_pt(image)
+        image = normalize(image)  # to [-1, 1]
+
+        # 4. Add temporal dimension
+        # After numpy_to_pt: [1, C, H, W] (batch already added by pil_to_numpy)
+        # Add T dimension: [1, C, H, W] -> [1, C, 1, H, W] = [B, C, T, H, W]
+        image = image.unsqueeze(2)
+        image = image.to(get_local_torch_device(), dtype=torch.float32)
+
+        # 5. Encode via VAE
+        self.vae = self.vae.to(get_local_torch_device())
+
+        # Setup VAE precision
+        vae_dtype = PRECISION_TO_TYPE[
+            fastvideo_args.pipeline_config.vae_precision]
+        vae_autocast_enabled = (
+            vae_dtype != torch.float32) and not fastvideo_args.disable_autocast
+
+        with torch.autocast(device_type="cuda",
+                            dtype=vae_dtype,
+                            enabled=vae_autocast_enabled):
+            if fastvideo_args.pipeline_config.vae_tiling:
+                self.vae.enable_tiling()
+
+            if not vae_autocast_enabled:
+                image = image.to(vae_dtype)
+
+            with torch.no_grad():
+                encoder_output = self.vae.encode(image)
+                latent = self.retrieve_latents(encoder_output, batch.generator)
+
+        # 6. Apply LongCat-specific normalization
+        # Formula: (latents - mean) / std
+        latent = self.normalize_latents(latent)
+
+        # 7. Calculate num_cond_latents
+        # Formula: 1 + (num_cond_frames - 1) // vae_temporal_scale
+        # For single image (num_cond_frames=1): always 1 latent frame
+        num_cond_frames = 1  # Single image
+        vae_temporal_scale = self.vae.config.scale_factor_temporal
+        batch.num_cond_latents = 1 + (num_cond_frames - 1) // vae_temporal_scale
+
+        # 8. Store in batch
+        batch.image_latent = latent
+        batch.num_cond_frames = 1
+
+        logger.info(
+            "I2V: Encoded image to latent shape %s, num_cond_latents=%s",
+            latent.shape, batch.num_cond_latents)
+
+        # Offload VAE if needed
+        if fastvideo_args.vae_cpu_offload:
+            self.vae.to("cpu")
+
+        return batch
+
+    def retrieve_latents(self, encoder_output: object,
+                         generator: torch.Generator | None) -> torch.Tensor:
+        """Sample from VAE posterior."""
+        # WAN VAE returns an object with .sample() method
+        if hasattr(encoder_output, 'sample'):
+            return encoder_output.sample(generator)
+        elif hasattr(encoder_output, 'latent_dist'):
+            return encoder_output.latent_dist.sample(generator)
+        elif hasattr(encoder_output, 'latents'):
+            return encoder_output.latents
+        else:
+            raise AttributeError("Could not access latents from encoder output")
+
+    def normalize_latents(self, latents: torch.Tensor) -> torch.Tensor:
+        """
+        Apply LongCat-specific latent normalization.
+        
+        Formula: (latents - mean) / std
+        
+        This matches the original LongCat implementation and is DIFFERENT
+        from standard VAE scaling (which uses scaling_factor).
+        """
+        if not hasattr(self.vae.config, 'latents_mean') or not hasattr(
+                self.vae.config, 'latents_std'):
+            raise ValueError(
+                "VAE config must have 'latents_mean' and 'latents_std' "
+                "for LongCat normalization")
+
+        latents_mean = torch.tensor(self.vae.config.latents_mean).view(
+            1, self.vae.config.z_dim, 1, 1, 1).to(latents.device, latents.dtype)
+
+        latents_std = torch.tensor(self.vae.config.latents_std).view(
+            1, self.vae.config.z_dim, 1, 1, 1).to(latents.device, latents.dtype)
+
+        return (latents - latents_mean) / latents_std

fastvideo/pipelines/stages/longcat_kv_cache_init.py

@@ -0,0 +1,123 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat KV Cache Initialization Stage for Video Continuation (VC).
+
+This stage pre-computes K/V cache for conditioning frames.
+
+"""
+
+import torch
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.forward_context import set_forward_context
+from fastvideo.logger import init_logger
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.stages.base import PipelineStage
+
+logger = init_logger(__name__)
+
+
+class LongCatKVCacheInitStage(PipelineStage):
+    """
+    Pre-compute KV cache for conditioning frames.
+
+    After this stage:
+    - batch.kv_cache_dict contains {block_idx: (k, v)}
+    - batch.cond_latents contains the conditioning latents
+    - batch.latents contains ONLY noise latents
+    """
+
+    def __init__(self, transformer):
+        super().__init__()
+        self.transformer = transformer
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        """Initialize KV cache from conditioning latents."""
+
+        # Check if KV cache is enabled
+        use_kv_cache = getattr(fastvideo_args.pipeline_config, 'use_kv_cache',
+                               True)
+        if not use_kv_cache:
+            batch.kv_cache_dict = {}
+            batch.use_kv_cache = False
+            logger.info("KV cache disabled, skipping initialization")
+            return batch
+
+        batch.use_kv_cache = True
+        offload_kv_cache = getattr(fastvideo_args.pipeline_config,
+                                   'offload_kv_cache', False)
+
+        # Get conditioning latents
+        num_cond_latents = batch.num_cond_latents
+        if num_cond_latents <= 0:
+            batch.kv_cache_dict = {}
+            logger.warning("num_cond_latents <= 0, skipping KV cache init")
+            return batch
+
+        # Extract conditioning latents
+        cond_latents = batch.latents[:, :, :num_cond_latents].clone()
+
+        logger.info(
+            "Initializing KV cache for %d conditioning latents, shape: %s",
+            num_cond_latents, cond_latents.shape)
+
+        # Timestep = 0 for conditioning (they are "clean")
+        B = cond_latents.shape[0]
+        T_cond = cond_latents.shape[2]
+        timestep = torch.zeros(B,
+                               T_cond,
+                               device=cond_latents.device,
+                               dtype=cond_latents.dtype)
+
+        # Empty prompt embeddings (cross-attn will be skipped)
+        max_seq_len = 512
+        # Get caption dimension from transformer config
+        caption_dim = self.transformer.config.caption_channels
+        empty_embeds = torch.zeros(B,
+                                   max_seq_len,
+                                   caption_dim,
+                                   device=cond_latents.device,
+                                   dtype=cond_latents.dtype)
+
+        # Get transformer dtype
+        if hasattr(self.transformer, 'module'):
+            transformer_dtype = next(self.transformer.module.parameters()).dtype
+        else:
+            transformer_dtype = next(self.transformer.parameters()).dtype
+
+        # Run transformer with return_kv=True, skip_crs_attn=True
+        with (
+                torch.no_grad(),
+                set_forward_context(
+                    current_timestep=0,
+                    attn_metadata=None,
+                    forward_batch=batch,
+                ),
+                torch.autocast(device_type='cuda', dtype=transformer_dtype),
+        ):
+            _, kv_cache_dict = self.transformer(
+                hidden_states=cond_latents.to(transformer_dtype),
+                encoder_hidden_states=empty_embeds.to(transformer_dtype),
+                timestep=timestep.to(transformer_dtype),
+                return_kv=True,
+                skip_crs_attn=True,
+                offload_kv_cache=offload_kv_cache,
+            )
+
+        # Store cache and save cond_latents for later concatenation
+        batch.kv_cache_dict = kv_cache_dict
+        batch.cond_latents = cond_latents
+
+        # Remove conditioning latents from main latents
+        # After this, batch.latents contains ONLY noise frames
+        batch.latents = batch.latents[:, :, num_cond_latents:]
+
+        logger.info(
+            "KV cache initialized: %d blocks, offload=%s, remaining latents shape: %s",
+            len(kv_cache_dict), offload_kv_cache, batch.latents.shape)
+
+        return batch

fastvideo/pipelines/stages/longcat_refine_init.py

@@ -257,12 +257,16 @@ class LongCatRefineInitStage(PipelineStage):
                     num_cond_frames_added, num_noise_frames_added,
                     new_num_frames)
 
-        # VAE encode
-        logger.info("Encoding stage1 video with VAE...")
+        # VAE encode with tiling for memory efficiency
+        logger.info("Encoding stage1 video with VAE (tiling enabled)...")
         vae_dtype = next(self.vae.parameters()).dtype
         vae_device = next(self.vae.parameters()).device
         video_up = video_up.to(dtype=vae_dtype, device=vae_device)
 
+        # Enable tiling for large video encoding
+        if hasattr(self.vae, 'enable_tiling'):
+            self.vae.enable_tiling()
+
         with torch.no_grad():
             latent_dist = self.vae.encode(video_up)
             # Extract tensor from latent distribution
@@ -301,10 +305,14 @@ class LongCatRefineInitStage(PipelineStage):
 
         logger.info("Applied t_thresh=%s noise mixing", t_thresh)
 
-        # Store in batch
-        batch.latents = latent_up.to(dtype)
+        # Store in batch - ensure correct dtype and device
+        # The latents need to be on the same device as the transformer (CUDA)
+        target_device = batch.prompt_embeds[0].device
+        batch.latents = latent_up.to(device=target_device, dtype=dtype)
         batch.raw_latent_shape = latent_up.shape
 
+        logger.info("Latents device: %s, dtype: %s", batch.latents.device,
+                    batch.latents.dtype)
         logger.info("LongCat refinement initialization complete")
 
         return batch

fastvideo/pipelines/stages/longcat_vc_denoising.py

@@ -0,0 +1,217 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat VC Denoising Stage with KV cache support.
+
+This stage extends the I2V denoising stage to support:
+1. KV cache for conditioning frames
+2. Video continuation with multiple conditioning frames
+"""
+
+import time
+import torch
+from tqdm import tqdm
+
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.forward_context import set_forward_context
+from fastvideo.logger import init_logger
+from fastvideo.models.loader.component_loader import TransformerLoader
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.stages.longcat_denoising import LongCatDenoisingStage
+
+logger = init_logger(__name__)
+
+
+class LongCatVCDenoisingStage(LongCatDenoisingStage):
+    """
+    LongCat denoising with Video Continuation and KV cache support.
+    
+    Key differences from I2V denoising:
+    - Supports KV cache (reuses cached K/V from conditioning frames)
+    - Handles larger num_cond_latents
+    - Concatenates conditioning latents back after denoising
+    
+    When use_kv_cache=True:
+    - batch.latents contains ONLY noise frames (cond removed by KV cache init)
+    - batch.kv_cache_dict contains cached K/V
+    - batch.cond_latents contains conditioning latents for post-concat
+    
+    When use_kv_cache=False:
+    - batch.latents contains ALL frames (cond + noise)
+    - Timestep masking: timestep[:, :num_cond_latents] = 0
+    - Selective denoising: only update noise frames
+    """
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        """Run denoising loop with VC conditioning and optional KV cache."""
+
+        # Load transformer if needed
+        if not fastvideo_args.model_loaded["transformer"]:
+            loader = TransformerLoader()
+            self.transformer = loader.load(
+                fastvideo_args.model_paths["transformer"], fastvideo_args)
+            fastvideo_args.model_loaded["transformer"] = True
+
+        # Setup
+        target_dtype = torch.bfloat16
+        autocast_enabled = (target_dtype != torch.float32
+                            ) and not fastvideo_args.disable_autocast
+
+        latents = batch.latents
+        timesteps = batch.timesteps
+        prompt_embeds = batch.prompt_embeds[0]
+        prompt_attention_mask = (batch.prompt_attention_mask[0]
+                                 if batch.prompt_attention_mask else None)
+        guidance_scale = batch.guidance_scale
+        do_classifier_free_guidance = batch.do_classifier_free_guidance
+
+        # Get VC-specific parameters
+        num_cond_latents = getattr(batch, 'num_cond_latents', 0)
+        use_kv_cache = getattr(batch, 'use_kv_cache', False)
+        kv_cache_dict = getattr(batch, 'kv_cache_dict', {})
+
+        logger.info(
+            "VC Denoising: num_cond_latents=%d, use_kv_cache=%s, latent_shape=%s",
+            num_cond_latents, use_kv_cache, latents.shape)
+
+        # Prepare negative prompts for CFG
+        if do_classifier_free_guidance:
+            negative_prompt_embeds = batch.negative_prompt_embeds[0]
+            negative_prompt_attention_mask = (batch.negative_attention_mask[0]
+                                              if batch.negative_attention_mask
+                                              else None)
+
+            prompt_embeds_combined = torch.cat(
+                [negative_prompt_embeds, prompt_embeds], dim=0)
+            if prompt_attention_mask is not None:
+                prompt_attention_mask_combined = torch.cat(
+                    [negative_prompt_attention_mask, prompt_attention_mask],
+                    dim=0)
+            else:
+                prompt_attention_mask_combined = None
+        else:
+            prompt_embeds_combined = prompt_embeds
+            prompt_attention_mask_combined = prompt_attention_mask
+
+        # Denoising loop
+        num_inference_steps = len(timesteps)
+        step_times = []
+
+        with tqdm(total=num_inference_steps,
+                  desc="VC Denoising") as progress_bar:
+            for i, t in enumerate(timesteps):
+                step_start = time.time()
+
+                # 1. Expand latents for CFG
+                if do_classifier_free_guidance:
+                    latent_model_input = torch.cat([latents] * 2)
+                else:
+                    latent_model_input = latents
+
+                latent_model_input = latent_model_input.to(target_dtype)
+
+                # 2. Expand timestep to match batch size
+                timestep = t.expand(
+                    latent_model_input.shape[0]).to(target_dtype)
+
+                # 3. Expand timestep to temporal dimension
+                timestep = timestep.unsqueeze(-1).repeat(
+                    1, latent_model_input.shape[2])
+
+                # 4. Timestep masking (only when NOT using KV cache)
+                if not use_kv_cache and num_cond_latents > 0:
+                    timestep[:, :num_cond_latents] = 0
+
+                # 5. Prepare transformer kwargs
+                # IMPORTANT: num_cond_latents is ALWAYS passed - needed for RoPE position offset
+                transformer_kwargs = {
+                    'num_cond_latents': num_cond_latents,
+                }
+                if use_kv_cache:
+                    transformer_kwargs['kv_cache_dict'] = kv_cache_dict
+
+                # 6. Run transformer
+                batch.is_cfg_negative = False
+                with set_forward_context(
+                        current_timestep=i,
+                        attn_metadata=None,
+                        forward_batch=batch,
+                ), torch.autocast(device_type='cuda',
+                                  dtype=target_dtype,
+                                  enabled=autocast_enabled):
+                    noise_pred = self.transformer(
+                        hidden_states=latent_model_input,
+                        encoder_hidden_states=prompt_embeds_combined,
+                        timestep=timestep,
+                        encoder_attention_mask=prompt_attention_mask_combined,
+                        **transformer_kwargs,
+                    )
+
+                # 7. Apply CFG with optimized scale (CFG-zero)
+                if do_classifier_free_guidance:
+                    noise_pred_uncond, noise_pred_cond = noise_pred.chunk(2)
+
+                    B = noise_pred_cond.shape[0]
+                    positive = noise_pred_cond.reshape(B, -1)
+                    negative = noise_pred_uncond.reshape(B, -1)
+
+                    st_star = self.optimized_scale(positive, negative)
+                    st_star = st_star.view(B, 1, 1, 1, 1)
+
+                    noise_pred = (
+                        noise_pred_uncond * st_star + guidance_scale *
+                        (noise_pred_cond - noise_pred_uncond * st_star))
+
+                # 8. Negate for flow matching scheduler
+                noise_pred = -noise_pred
+
+                # 9. Scheduler step
+                if use_kv_cache:
+                    # All latents are noise frames (conditioning is in cache)
+                    latents = self.scheduler.step(noise_pred,
+                                                  t,
+                                                  latents,
+                                                  return_dict=False)[0]
+                else:
+                    # Only update noise frames (skip conditioning)
+                    if num_cond_latents > 0:
+                        latents[:, :, num_cond_latents:] = self.scheduler.step(
+                            noise_pred[:, :, num_cond_latents:],
+                            t,
+                            latents[:, :, num_cond_latents:],
+                            return_dict=False,
+                        )[0]
+                    else:
+                        latents = self.scheduler.step(noise_pred,
+                                                      t,
+                                                      latents,
+                                                      return_dict=False)[0]
+
+                step_time = time.time() - step_start
+                step_times.append(step_time)
+
+                # Log timing for first few steps
+                if i < 3:
+                    logger.info("Step %d: %.2fs", i, step_time)
+
+                progress_bar.update()
+
+        # 10. If using KV cache, concatenate conditioning latents back
+        if use_kv_cache and hasattr(
+                batch, 'cond_latents') and batch.cond_latents is not None:
+            latents = torch.cat([batch.cond_latents, latents], dim=2)
+            logger.info(
+                "Concatenated conditioning latents back, final shape: %s",
+                latents.shape)
+
+        # Log average timing
+        avg_time = sum(step_times) / len(step_times)
+        logger.info("Average step time: %.2fs (total: %.1fs)", avg_time,
+                    sum(step_times))
+
+        # Update batch with denoised latents
+        batch.latents = latents
+        return batch

fastvideo/pipelines/stages/longcat_video_vae_encoding.py

@@ -0,0 +1,180 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+LongCat Video VAE Encoding Stage for Video Continuation (VC) generation.
+
+This stage handles encoding multiple video frames to latent space with
+LongCat-specific normalization for VC conditioning.
+"""
+
+from typing import Any
+
+import PIL.Image
+import torch
+
+from fastvideo.distributed import get_local_torch_device
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.logger import init_logger
+from fastvideo.models.vision_utils import normalize, numpy_to_pt, pil_to_numpy, resize
+from fastvideo.pipelines.pipeline_batch_info import ForwardBatch
+from fastvideo.pipelines.stages.base import PipelineStage
+from fastvideo.utils import PRECISION_TO_TYPE
+
+logger = init_logger(__name__)
+
+
+class LongCatVideoVAEEncodingStage(PipelineStage):
+    """
+    Encode video frames to latent space for VC conditioning.
+    
+    This stage:
+    1. Loads video frames from path or uses provided frames
+    2. Takes the last num_cond_frames from the video
+    3. Preprocesses and stacks frames
+    4. Encodes via VAE to latent space
+    5. Applies LongCat-specific normalization
+    6. Calculates num_cond_latents
+    """
+
+    def __init__(self, vae):
+        super().__init__()
+        self.vae = vae
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        """Encode video frames to latent for VC conditioning."""
+
+        # Get video from batch - can be path, list of PIL images, or already loaded
+        video = getattr(batch, 'video_frames', None) or getattr(
+            batch, 'video_path', None)
+        num_cond_frames = getattr(batch, 'num_cond_frames',
+                                  13)  # Default 13 for VC
+
+        if video is None:
+            raise ValueError(
+                "video_frames or video_path must be provided for VC")
+
+        # Load video if path
+        if isinstance(video, str):
+            from diffusers.utils import load_video
+            video = load_video(video)
+            logger.info("Loaded video from path: %d frames", len(video))
+
+        # Take last num_cond_frames
+        if len(video) > num_cond_frames:
+            video = video[-num_cond_frames:]
+            logger.info("Using last %d frames for conditioning",
+                        num_cond_frames)
+        elif len(video) < num_cond_frames:
+            logger.warning(
+                "Video has only %d frames, less than num_cond_frames=%d",
+                len(video), num_cond_frames)
+            num_cond_frames = len(video)
+
+        # Get target dimensions
+        height = batch.height
+        width = batch.width
+
+        if height is None or width is None:
+            raise ValueError("height and width must be set for VC")
+
+        # Preprocess and stack frames
+        processed_frames = []
+        for frame in video:
+            if not isinstance(frame, PIL.Image.Image):
+                raise TypeError(f"Frame must be PIL.Image, got {type(frame)}")
+
+            frame = resize(frame, height, width, resize_mode="default")
+            frame = pil_to_numpy(frame)  # Returns [1, H, W, C] then converted
+            frame = numpy_to_pt(frame)  # Returns [1, C, H, W]
+            frame = normalize(frame)  # to [-1, 1]
+            processed_frames.append(frame)
+
+        # Stack frames: [num_frames, C, H, W] -> [1, C, T, H, W]
+        video_tensor = torch.cat(processed_frames, dim=0)  # [T, C, H, W]
+        video_tensor = video_tensor.permute(1, 0, 2,
+                                            3).unsqueeze(0)  # [1, C, T, H, W]
+        video_tensor = video_tensor.to(get_local_torch_device(),
+                                       dtype=torch.float32)
+
+        logger.info("VC: Preprocessed video tensor shape: %s",
+                    video_tensor.shape)
+
+        # Encode via VAE
+        self.vae = self.vae.to(get_local_torch_device())
+
+        # Setup VAE precision
+        vae_dtype = PRECISION_TO_TYPE[
+            fastvideo_args.pipeline_config.vae_precision]
+        vae_autocast_enabled = (
+            vae_dtype != torch.float32) and not fastvideo_args.disable_autocast
+
+        with torch.autocast(device_type="cuda",
+                            dtype=vae_dtype,
+                            enabled=vae_autocast_enabled):
+            if fastvideo_args.pipeline_config.vae_tiling:
+                self.vae.enable_tiling()
+
+            if not vae_autocast_enabled:
+                video_tensor = video_tensor.to(vae_dtype)
+
+            with torch.no_grad():
+                encoder_output = self.vae.encode(video_tensor)
+                latent = self.retrieve_latents(encoder_output, batch.generator)
+
+        # Apply LongCat-specific normalization
+        latent = self.normalize_latents(latent)
+
+        # Calculate num_cond_latents
+        # Formula: 1 + (num_cond_frames - 1) // vae_temporal_scale
+        vae_temporal_scale = self.vae.config.scale_factor_temporal
+        num_cond_latents = 1 + (num_cond_frames - 1) // vae_temporal_scale
+
+        # Store in batch
+        batch.video_latent = latent
+        batch.num_cond_frames = num_cond_frames
+        batch.num_cond_latents = num_cond_latents
+
+        logger.info(
+            "VC: Encoded %d frames to latent shape %s, num_cond_latents=%d",
+            num_cond_frames, latent.shape, num_cond_latents)
+
+        # Offload VAE if needed
+        if fastvideo_args.vae_cpu_offload:
+            self.vae.to("cpu")
+
+        return batch
+
+    def retrieve_latents(self, encoder_output: Any,
+                         generator: torch.Generator | None) -> torch.Tensor:
+        """Sample from VAE posterior."""
+        if hasattr(encoder_output, 'sample'):
+            return encoder_output.sample(generator)
+        elif hasattr(encoder_output, 'latent_dist'):
+            return encoder_output.latent_dist.sample(generator)
+        elif hasattr(encoder_output, 'latents'):
+            return encoder_output.latents
+        else:
+            raise AttributeError("Could not access latents from encoder output")
+
+    def normalize_latents(self, latents: torch.Tensor) -> torch.Tensor:
+        """
+        Apply LongCat-specific latent normalization.
+        
+        Formula: (latents - mean) / std
+        """
+        if not hasattr(self.vae.config, 'latents_mean') or not hasattr(
+                self.vae.config, 'latents_std'):
+            raise ValueError(
+                "VAE config must have 'latents_mean' and 'latents_std' "
+                "for LongCat normalization")
+
+        latents_mean = torch.tensor(self.vae.config.latents_mean).view(
+            1, self.vae.config.z_dim, 1, 1, 1).to(latents.device, latents.dtype)
+
+        latents_std = torch.tensor(self.vae.config.latents_std).view(
+            1, self.vae.config.z_dim, 1, 1, 1).to(latents.device, latents.dtype)
+
+        return (latents - latents_mean) / latents_std

fastvideo/pipelines/stages/text_encoding.py

@@ -328,3 +328,69 @@ class TextEncodingStage(PipelineStage):
             lambda x: not batch.do_classifier_free_guidance or V.
             list_of_tensors_with_min_dims(x, 2))
         return result
+
+
+class Cosmos25TextEncodingStage(PipelineStage):
+    """Cosmos 2.5 text encoding stage.
+
+    Cosmos 2.5 uses Reason1 (Qwen2.5-VL) and relies on the encoder's
+    `compute_text_embeddings_online()`.
+    """
+
+    def __init__(self, text_encoder) -> None:
+        super().__init__()
+        self.text_encoder = text_encoder
+
+    @torch.no_grad()
+    def forward(self, batch: ForwardBatch,
+                fastvideo_args: FastVideoArgs) -> ForwardBatch:
+        assert batch.prompt is not None
+        prompts = [batch.prompt] if isinstance(batch.prompt,
+                                               str) else batch.prompt
+
+        encoder = self.text_encoder
+        if not hasattr(encoder, "compute_text_embeddings_online"):
+            raise RuntimeError(
+                "Cosmos25TextEncodingStage requires text_encoder.compute_text_embeddings_online()"
+            )
+
+        with set_forward_context(current_timestep=0, attn_metadata=None):
+            prompt_embeds = encoder.compute_text_embeddings_online(
+                {"text": prompts}, "text")
+
+        batch.prompt_embeds = [prompt_embeds]
+
+        if batch.do_classifier_free_guidance:
+            neg = batch.negative_prompt
+            neg_prompts = ([neg] *
+                           len(prompts)) if isinstance(neg, str) else neg
+            with set_forward_context(current_timestep=0, attn_metadata=None):
+                neg_embeds = encoder.compute_text_embeddings_online(
+                    {"text": neg_prompts}, "text")
+            batch.negative_prompt_embeds = [neg_embeds]
+        else:
+            batch.negative_prompt_embeds = []
+
+        return batch
+
+    def verify_input(self, batch: ForwardBatch,
+                     fastvideo_args: FastVideoArgs) -> VerificationResult:
+        result = VerificationResult()
+        result.add_check("prompt", batch.prompt, V.string_or_list_strings)
+        result.add_check(
+            "negative_prompt",
+            batch.negative_prompt,
+            lambda x:
+            (not batch.do_classifier_free_guidance) or isinstance(x, str),
+        )
+        return result
+
+    def verify_output(self, batch: ForwardBatch,
+                      fastvideo_args: FastVideoArgs) -> VerificationResult:
+        result = VerificationResult()
+        result.add_check("prompt_embeds", batch.prompt_embeds,
+                         V.list_of_tensors_min_dims(2))
+        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/timestep_preparation.py

@@ -123,3 +123,32 @@ class TimestepPreparationStage(PipelineStage):
         result.add_check("timesteps", batch.timesteps,
                          [V.is_tensor, V.with_dims(1)])
         return result
+
+
+class Cosmos25TimestepPreparationStage(TimestepPreparationStage):
+    """Cosmos 2.5 timestep preparation with scheduler-specific kwargs."""
+
+    def forward(
+        self,
+        batch: ForwardBatch,
+        fastvideo_args: FastVideoArgs,
+    ) -> ForwardBatch:
+        scheduler = self.scheduler
+        device = get_local_torch_device()
+        num_inference_steps = batch.num_inference_steps
+
+        extra_kwargs: dict = {}
+        sig = inspect.signature(scheduler.set_timesteps)
+        if "shift" in sig.parameters:
+            extra_kwargs["shift"] = fastvideo_args.pipeline_config.flow_shift
+        # Prefer the canonical diffusers kwarg name if available.
+        if "use_karras_sigmas" in sig.parameters:
+            extra_kwargs["use_karras_sigmas"] = True
+        elif "use_kerras_sigma" in sig.parameters:
+            extra_kwargs["use_kerras_sigma"] = True
+
+        scheduler.set_timesteps(num_inference_steps,
+                                device=device,
+                                **extra_kwargs)
+        batch.timesteps = scheduler.timesteps
+        return batch