[docs] Update README with realtime demo announcement

Made-with: Cursor

.buildkite/pipeline.yml

@@ -2,46 +2,20 @@ env:
   IMAGE_VERSION: "py3.12-latest"
   BUILDKITE_CLEAN_CHECKOUT: true
 
-notify:
-  - github_commit_status:
-      context: "fastcheck-passed"
-    if: build.env("TEST_SCOPE") == "fastcheck" || build.env("TEST_SCOPE") == null
-  - github_commit_status:
-      context: "full-suite-passed"
-    if: build.env("TEST_SCOPE") == "full"
-
 steps:
-  # ============================================================
-    - label: ":dart: Direct Test (${TEST_TYPE})"
-      if: build.env("TEST_SCOPE") == "direct"
-      command: "timeout 90m .buildkite/scripts/pr_test.sh"
-      retry:
-        automatic:
-          - exit_status: 128
-            limit: 3
-          - exit_status: -1
-            limit: 2
-      agents:
-        queue: "default"
+  - label: "pre-commit"
+    command: ".buildkite/scripts/pre_commit.sh"
+    agents:
+      queue: "default"
 
-  # ============================================================
-  # Fastcheck: Runs on every PR (~10-15 min parallel)
-  # Core component validation: encoders, VAEs, transformers,
-  # CUDA kernels, and unit tests.
-  # ============================================================
-    - label: "Trigger Fastcheck"
-      if: build.env("TEST_SCOPE") == "fastcheck" || build.env("TEST_SCOPE") == null
-      retry:
-        automatic:
-          - exit_status: 128
-            limit: 3
-          - exit_status: -1
-            limit: 2
-      plugins:
-        - monorepo-diff#v1.4.0:
-            diff: 'git fetch origin "${BUILDKITE_PULL_REQUEST_BASE_BRANCH:-main}" && git diff --name-only "origin/${BUILDKITE_PULL_REQUEST_BASE_BRANCH:-main}...HEAD"'
-            watch:
-            - path:
+  - wait
+
+  - label: "Trigger Tests"
+    plugins:
+      - monorepo-diff#v1.4.0:
+          diff: 'git fetch origin "$BUILDKITE_PULL_REQUEST_BASE_BRANCH" && git diff --name-only origin/"$BUILDKITE_PULL_REQUEST_BASE_BRANCH"...HEAD'
+          watch:
+            - path: 
                 - "fastvideo/models/encoders/**"
                 - "fastvideo/models/loader/**"
                 - "fastvideo/tests/encoders/**"
@@ -49,7 +23,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 20m .buildkite/scripts/pr_test.sh"
-                label: ":microscope: Encoder Tests"
+                label: "Encoder Tests"
                 env:
                   - TEST_TYPE=encoder
                 agents:
@@ -62,7 +36,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 20m .buildkite/scripts/pr_test.sh"
-                label: ":microscope: VAE Tests"
+                label: "VAE Tests"
                 env:
                   - TEST_TYPE=vae
                 agents:
@@ -77,62 +51,18 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":microscope: Transformer Tests"
+                label: "Transformer Tests"
                 env:
                   - TEST_TYPE=transformer
                 agents:
                   queue: "default"
-            - path:
-                - "fastvideo-kernel/**"
-                - "pyproject.toml"
-                - "docker/Dockerfile.python3.12"
-              config:
-                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":microscope: Kernel Tests"
-                env:
-                  - TEST_TYPE=kernel_tests
-                agents:
-                  queue: "default"
-            - path:
-                - "fastvideo/**"
-                - ".buildkite/**"
-                - ".github/**"
-                - "pyproject.toml"
-                - "docker/Dockerfile.python3.12"
-              config:
-                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":microscope: Unit Tests"
-                env:
-                  - TEST_TYPE=unit_test
-                agents:
-                  queue: "default"
-
-  # ============================================================
-  # Full Suite: Runs when TEST_SCOPE=full
-  # Triggered by adding the 'ready' label (via ci-trigger-full-suite.yml)
-  # or on-demand via /test full slash command.
-  # Includes integration tests, SSIM regression, training pipelines,
-  # and performance benchmarks.
-  # ============================================================
-    - label: "Trigger Full Suite"
-      if: build.env("TEST_SCOPE") == "full"
-      retry:
-        automatic:
-          - exit_status: 128
-            limit: 3
-          - exit_status: -1
-            limit: 2
-      plugins:
-        - monorepo-diff#v1.4.0:
-            diff: 'git fetch origin "${BUILDKITE_PULL_REQUEST_BASE_BRANCH:-main}" && git diff --name-only "origin/${BUILDKITE_PULL_REQUEST_BASE_BRANCH:-main}...HEAD"'
-            watch:
-            - path:
+            - path: 
                 - "fastvideo/**/*.py"
                 - "pyproject.toml"
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 90m .buildkite/scripts/pr_test.sh"
-                label: ":bar_chart: SSIM Tests"
+                label: "SSIM Tests"
                 env:
                   - TEST_TYPE=ssim
                 agents:
@@ -147,7 +77,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 20m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: LoRA Inference Tests"
+                label: "LoRA Inference Tests"
                 env:
                   - TEST_TYPE=inference_lora
                 agents:
@@ -158,7 +88,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: Training Tests"
+                label: "Training Tests"
                 env:
                   - TEST_TYPE=training
                 agents:
@@ -169,7 +99,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: Distillation DMD Tests"
+                label: "Distillation DMDTests"
                 env:
                   - TEST_TYPE=distillation_dmd
                 agents:
@@ -181,7 +111,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: Self-Forcing Tests"
+                label: "Self-Forcing Tests"
                 env:
                   - TEST_TYPE=self_forcing
                 agents:
@@ -192,7 +122,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: LoRA Training Tests"
+                label: "LoRA Training Tests"
                 env:
                   - TEST_TYPE=training_lora
                 agents:
@@ -204,11 +134,20 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: Training Tests VSA"
+                label: "Training Tests VSA"
                 env:
                   - TEST_TYPE=training_vsa
                 agents:
                   queue: "default"
+            - path:
+                - "fastvideo-kernel/**"
+                - "pyproject.toml"
+                - "docker/Dockerfile.python3.12"
+              config:
+                command: "timeout 15m .buildkite/scripts/pr_test.sh"
+                label: "Kernel Tests"
+                env:
+                  - TEST_TYPE=kernel_tests
             - path:
                 - "fastvideo-kernel/**"
                 - "fastvideo/attention/backends/vmoba.py"
@@ -216,11 +155,22 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: Inference Tests VMoBA"
-                env:
+                label: "Inference Tests VMoBA"
+                env:  
                   - TEST_TYPE=inference_vmoba
                 agents:
                   queue: "default"
+            - path:
+                - "fastvideo/**"
+                - "pyproject.toml"
+                - "docker/Dockerfile.python3.12"
+              config:
+                command: "timeout 15m .buildkite/scripts/pr_test.sh"
+                label: "Unit Tests"
+                env:
+                  - TEST_TYPE=unit_test
+                agents:
+                  queue: "default"
             - path:
                 - "fastvideo/models/dits/**"
                 - "fastvideo/pipelines/**"
@@ -234,7 +184,7 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 30m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: Performance Tests"
+                label: "Performance Tests"
                 env:
                   - TEST_TYPE=performance
                 agents:
@@ -247,8 +197,19 @@ steps:
                 - "docker/Dockerfile.python3.12"
               config:
                 command: "timeout 30m .buildkite/scripts/pr_test.sh"
-                label: ":test_tube: API Server Tests"
+                label: "API Server Tests"
                 env:
                   - TEST_TYPE=api_server
                 agents:
                   queue: "default"
+            # - path:
+            #     - "scripts/lora_extraction/**"
+            #     - "pyproject.toml"
+            #     - "docker/Dockerfile.python3.12"
+            #   config:
+            #     command: "timeout 90m .buildkite/scripts/pr_test.sh"
+            #     label: "LoRA Extraction Tests"
+            #     env:
+            #       - TEST_TYPE=lora_extraction
+            #     agents:
+            #       queue: "default"

.buildkite/scripts/pr_test.sh

@@ -59,11 +59,7 @@ if [ -z "${TEST_TYPE:-}" ]; then
 fi
 log "Test type: $TEST_TYPE"
 
-EFFECTIVE_PR=${BUILDKITE_PULL_REQUEST:-false}
-if [ "$EFFECTIVE_PR" = "false" ] && [ -n "${PR_NUMBER:-}" ]; then
-    EFFECTIVE_PR=$PR_NUMBER
-fi
-MODAL_ENV="BUILDKITE_REPO=$BUILDKITE_REPO BUILDKITE_COMMIT=$BUILDKITE_COMMIT BUILDKITE_PULL_REQUEST=$EFFECTIVE_PR IMAGE_VERSION=$IMAGE_VERSION"
+MODAL_ENV="BUILDKITE_REPO=$BUILDKITE_REPO BUILDKITE_COMMIT=$BUILDKITE_COMMIT BUILDKITE_PULL_REQUEST=$BUILDKITE_PULL_REQUEST IMAGE_VERSION=$IMAGE_VERSION"
 
 case "$TEST_TYPE" in
     "encoder")

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,22 +1,3 @@
-<!--
-PR TITLE: Must start with a type tag, e.g.:
-  [feat] Add new model       [bugfix] Fix VAE tiling      [refactor] Restructure pipeline
-  [perf] Optimize kernel     [ci] Update tests             [docs] Add guide
-  [misc] Cleanup configs     [new-model] Port Flux2
-
-MERGE WORKFLOW:
-  1. Ensure pre-commit passes and you have at least 1 approval
-  2. Comment /merge (or add the "ready" label) to enter the Merge Queue
-  3. Full Test Suite runs automatically on a staging branch → auto-merge on success
-
-ON-DEMAND TESTING (write access required):
-  /test full       — Full Test Suite        /test ssim        — SSIM regression
-  /test training   — Training pipeline      /test encoder     — Encoder tests
-  /test transformer — Transformer tests     /test vae         — VAE tests
-  /test kernel     — CUDA kernel tests      /test unit        — Unit tests
-  See docs/contributing/pull_requests.md for all 17 test commands
--->
-
 ## Purpose
 
 <!-- What does this PR do? Link the related issue if applicable. -->

.github/mergify.yml

@@ -1,321 +0,0 @@
-merge_protections:
-  - name: PR merge requirements
-    if:
-      - base = main
-    success_conditions:
-      - "title~=(?i)^\\[(feat|feature|bugfix|fix|refactor|perf|ci|doc|docs|misc|chore|kernel|new.?model)\\]"
-      - check-success~=pre-commit
-      - check-success=fastcheck-passed
-
-pull_request_rules:
-
-  # ============================================================
-  # Type labels (from PR title prefix)
-  # ============================================================
-
-  - name: "label type: feat"
-    conditions:
-      - "title~=(?i)^\\[(feat|feature)\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: feat"]
-
-  - name: "label type: bugfix"
-    conditions:
-      - "title~=(?i)^\\[(bug)?fix\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: bugfix"]
-
-  - name: "label type: refactor"
-    conditions:
-      - "title~=(?i)^\\[refactor\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: refactor"]
-
-  - name: "label type: perf"
-    conditions:
-      - "title~=(?i)^\\[perf\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: perf"]
-
-  - name: "label type: ci"
-    conditions:
-      - "title~=(?i)^\\[ci\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: ci"]
-
-  - name: "label type: docs"
-    conditions:
-      - "title~=(?i)^\\[(doc|docs)\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: docs"]
-
-  - name: "label type: misc"
-    conditions:
-      - "title~=(?i)^\\[(misc|chore)\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: misc"]
-
-  - name: "label type: new-model"
-    conditions:
-      - "title~=(?i)^\\[new.?model\\]"
-      - -closed
-    actions:
-      label:
-        add: ["type: new-model"]
-
-  # ============================================================
-  # Scope labels (from changed files)
-  # ============================================================
-
-  - name: "label scope: training"
-    conditions:
-      - or:
-          - files~=^fastvideo/train/
-          - files~=^fastvideo/training/
-          - files~=^fastvideo/distillation/
-          - files~=^examples/train/
-          - files~=^examples/training/
-          - files~=^examples/distill/
-      - -closed
-    actions:
-      label:
-        add: ["scope: training"]
-
-  - name: "label scope: inference"
-    conditions:
-      - or:
-          - files~=^fastvideo/pipelines/basic/
-          - files~=^fastvideo/pipelines/stages/
-          - files~=^fastvideo/pipelines/samplers/
-          - files~=^fastvideo/entrypoints/
-          - files~=^fastvideo/worker/
-          - files~=^fastvideo/configs/sample/
-          - files~=^fastvideo/configs/pipelines/
-          - files~=^examples/inference/
-      - -closed
-    actions:
-      label:
-        add: ["scope: inference"]
-
-  - name: "label scope: attention"
-    conditions:
-      - files~=^fastvideo/attention/
-      - -closed
-    actions:
-      label:
-        add: ["scope: attention"]
-
-  - name: "label scope: kernel"
-    conditions:
-      - or:
-          - files~=^fastvideo-kernel/
-          - files~=^csrc/
-      - -closed
-    actions:
-      label:
-        add: ["scope: kernel"]
-
-  - name: "label scope: data"
-    conditions:
-      - or:
-          - files~=^fastvideo/dataset/
-          - files~=^fastvideo/pipelines/preprocess/
-          - files~=^examples/preprocessing/
-      - -closed
-    actions:
-      label:
-        add: ["scope: data"]
-
-  - name: "label scope: infra"
-    conditions:
-      - or:
-          - files~=^\.github/
-          - files~=^\.buildkite/
-          - files~=^fastvideo/tests/
-          - files~=^docker/
-      - -closed
-    actions:
-      label:
-        add: ["scope: infra"]
-
-  - name: "label scope: distributed"
-    conditions:
-      - files~=^fastvideo/distributed/
-      - -closed
-    actions:
-      label:
-        add: ["scope: distributed"]
-
-  - name: "label scope: docs"
-    conditions:
-      - files~=^docs/
-      - -closed
-    actions:
-      label:
-        add: ["scope: docs"]
-
-  - name: "label scope: ui"
-    conditions:
-      - files~=^ui/
-      - -closed
-    actions:
-      label:
-        add: ["scope: ui"]
-
-  - name: "label scope: model"
-    conditions:
-      - or:
-          - files~=^fastvideo/models/
-          - files~=^fastvideo/layers/
-          - files~=^fastvideo/configs/models/
-      - -closed
-    actions:
-      label:
-        add: ["scope: model"]
-
-  # ============================================================
-  # Pre-commit failure help comment
-  # ============================================================
-
-  - name: comment on pre-commit failure
-    conditions:
-      - check-failure~=pre-commit
-      - -closed
-    actions:
-      comment:
-        message: |
-          ## Pre-commit checks failed
-
-          Hi @{{author}}, the pre-commit checks have failed. To fix them locally:
-
-          ```bash
-          # Install pre-commit if you haven't already
-          uv pip install pre-commit
-          pre-commit install
-
-          # Run all checks and auto-fix what's possible
-          pre-commit run --all-files
-          ```
-
-          Common fixes:
-          - **yapf**: `yapf -i <file>` (formatting)
-          - **ruff**: `ruff check --fix <file>` (linting)
-          - **codespell**: `codespell --write-changes <file>` (spelling)
-
-          After fixing, commit and push the changes. The checks will re-run automatically.
-
-          For future commits, `pre-commit` will run automatically on changed files before each commit.
-
-
-  # ============================================================
-  # Merge conflict detection
-  # ============================================================
-
-  - name: label conflicting PRs
-    conditions:
-      - conflict
-      - -closed
-      - label!=stale
-    actions:
-      label:
-        add: [needs-rebase]
-      comment:
-        message: |
-          This PR has merge conflicts with the base branch. Please rebase:
-
-          ```bash
-          git fetch origin main
-          git rebase origin/main
-          # Resolve any conflicts, then:
-          git push --force-with-lease
-          ```
-
-  - name: remove conflict label when resolved
-    conditions:
-      - -conflict
-      - -closed
-      - label=needs-rebase
-    actions:
-      label:
-        remove: [needs-rebase]
-
-  # ============================================================
-  # Auto-merge and auto-rebase
-  # ============================================================
-
-  - name: auto-merge when ready and all checks pass
-    conditions:
-      - label=ready
-      - "title~=(?i)^\\[(feat|feature|bugfix|fix|refactor|perf|ci|doc|docs|misc|chore|kernel|new.?model)\\]"
-      - "#approved-reviews-by>=1"
-      - check-success~=pre-commit
-      - check-success=fastcheck-passed
-      - check-success=full-suite-passed
-      - -conflict
-      - -closed
-      - -draft
-    actions:
-      merge:
-        method: squash
-
-  - name: auto-rebase when ready and Full Suite passed
-    conditions:
-      - label=ready
-      - "#approved-reviews-by>=1"
-      - check-success=full-suite-passed
-      - -conflict
-      - -closed
-      - -draft
-    actions:
-      rebase: {}
-
-  - name: remove ready label on Full Suite failure
-    conditions:
-      - label=ready
-      - check-failure=full-suite-passed
-    actions:
-      label:
-        remove: [ready]
-
-  # ============================================================
-  # PR title format help
-  # ============================================================
-
-  - name: comment on invalid PR title format
-    conditions:
-      - -closed
-      - -draft
-      - "-title~=(?i)^\\[(feat|feature|bugfix|fix|refactor|perf|ci|doc|docs|misc|chore|kernel|new.?model)\\]"
-    actions:
-      comment:
-        message: |
-          ## ⚠️ PR title format required
-
-          Your PR title must start with a type tag in brackets. Examples:
-          - `[feat] Add new model support`
-          - `[bugfix] Fix VAE tiling corruption`
-          - `[refactor] Restructure training pipeline`
-          - `[perf] Optimize attention kernel`
-          - `[ci] Update test infrastructure`
-          - `[docs] Add inference guide`
-          - `[misc] Clean up configs`
-          - `[new-model] Port Flux2 to FastVideo`
-
-          Valid tags: `feat`, `feature`, `bugfix`, `fix`, `refactor`, `perf`, `ci`, `doc`, `docs`, `misc`, `chore`, `kernel`, `new-model`
-
-          Please update your PR title and the merge protection check will pass automatically.
-

.github/scripts/runpod_api.py

@@ -0,0 +1,249 @@
+import argparse
+import json
+import os
+import subprocess
+import sys
+import time
+
+import requests
+
+
+def parse_arguments():
+    """Parse command line arguments"""
+    parser = argparse.ArgumentParser(description='Run tests on RunPod GPU')
+    parser.add_argument('--gpu-type', type=str, help='GPU type to use')
+    parser.add_argument('--gpu-count',
+                        type=int,
+                        help='Number of GPUs to use',
+                        default=1)
+    parser.add_argument('--test-command', type=str, help='Test command to run')
+    parser.add_argument('--disk-size',
+                        type=int,
+                        default=20,
+                        help='Container disk size in GB (default: 20)')
+    parser.add_argument('--volume-size',
+                        type=int,
+                        default=20,
+                        help='Persistent volume size in GB (default: 20)')
+    parser.add_argument(
+        '--image',
+        type=str,
+        required=True,
+        help='Docker image to use')
+    return parser.parse_args()
+
+
+args = parse_arguments()
+API_KEY = os.environ['RUNPOD_API_KEY']
+RUN_ID = os.environ['GITHUB_RUN_ID']
+JOB_ID = os.environ['JOB_ID']
+PODS_API = "https://rest.runpod.io/v1/pods"
+HEADERS = {
+    "Content-Type": "application/json",
+    "Authorization": f"Bearer {API_KEY}"
+}
+
+
+def create_pod():
+    """Create a RunPod instance"""
+    # Ensure image name is lowercase (Docker requirement)
+    image_name = args.image.lower()
+    print(f"Using specified image: {image_name}")
+
+    docker_start_cmd = [
+        "bash", 
+        "-c", 
+        "apt update;DEBIAN_FRONTEND=noninteractive apt-get install openssh-server -y;mkdir -p ~/.ssh;cd $_;chmod 700 ~/.ssh;echo \"$PUBLIC_KEY\" >> authorized_keys;chmod 700 authorized_keys;service ssh start;sleep infinity"
+    ]
+    
+    print(f"Creating RunPod instance with GPU: {args.gpu_type}...")
+    payload = {
+        "name": f"fastvideo-{JOB_ID}-{RUN_ID}",
+        "containerDiskInGb": args.disk_size,
+        "volumeInGb": args.volume_size,
+        "gpuTypeIds": [args.gpu_type],
+        "gpuCount": args.gpu_count,
+        "imageName": image_name,
+        "allowedCudaVersions": ["12.4"],
+        "dockerStartCmd": docker_start_cmd
+    }
+
+    response = requests.post(PODS_API, headers=HEADERS, json=payload)
+    response_data = response.json()
+    print(f"Response: {json.dumps(response_data, indent=2)}")
+
+    return response_data["id"]
+
+
+def wait_for_pod(pod_id):
+    """Wait for pod to be in RUNNING state and fully ready with SSH access"""
+    print("Waiting for RunPod to be ready...")
+
+    # First wait for RUNNING status
+    max_attempts = 10
+    attempts = 0
+    while attempts < max_attempts:
+        response = requests.get(f"{PODS_API}/{pod_id}", headers=HEADERS)
+        pod_data = response.json()
+        status = pod_data["desiredStatus"]
+
+        if status == "RUNNING":
+            print("RunPod is running! Now waiting for ports to be assigned...")
+            break
+
+        print(
+            f"Current status: {status}, waiting... (attempt {attempts+1}/{max_attempts})"
+        )
+        time.sleep(2)
+        attempts += 1
+
+    if attempts >= max_attempts:
+        raise TimeoutError(
+            "Timed out waiting for RunPod to reach RUNNING state")
+
+    # Wait for ports to be assigned
+    max_attempts = 50
+    attempts = 0
+    while attempts < max_attempts:
+        response = requests.get(f"{PODS_API}/{pod_id}", headers=HEADERS)
+        pod_data = response.json()
+        port_mappings = pod_data.get("portMappings")
+
+        if (port_mappings is not None and "22" in port_mappings
+                and pod_data.get("publicIp", "") != ""):
+            print("RunPod is ready with SSH access!")
+            print(f"SSH IP: {pod_data['publicIp']}")
+            print(f"SSH Port: {port_mappings['22']}")
+            break
+
+        print(
+            f"Waiting for SSH port and public IP to be available... (attempt {attempts+1}/{max_attempts})"
+        )
+        time.sleep(20)
+        attempts += 1
+
+    if attempts >= max_attempts:
+        raise TimeoutError("Timed out waiting for RunPod SSH access")
+
+
+def execute_command(pod_id):
+    """Execute command on the pod via SSH using system SSH client"""
+    print(f"Running command: {args.test_command}")
+
+    response = requests.get(f"{PODS_API}/{pod_id}", headers=HEADERS)
+    pod_data = response.json()
+    ssh_ip = pod_data["publicIp"]
+    ssh_port = pod_data["portMappings"]["22"]
+
+    # Copy the repository to the pod using scp
+    repo_dir = os.path.abspath(os.getcwd())
+    repo_name = os.path.basename(repo_dir)
+
+    print(f"Copying repository from {repo_dir} to RunPod...")
+
+    tar_command = [
+        "tar", "-czf", "/tmp/repo.tar.gz", "-C",
+        os.path.dirname(repo_dir), repo_name
+    ]
+    subprocess.run(tar_command, check=True)
+
+    # Copy the tarball to the pod
+    scp_command = [
+        "scp", "-o", "StrictHostKeyChecking=no", "-o",
+        "UserKnownHostsFile=/dev/null", "-o", "ServerAliveInterval=60", "-o",
+        "ServerAliveCountMax=10", "-P",
+        str(ssh_port), "/tmp/repo.tar.gz", f"root@{ssh_ip}:/tmp/"
+    ]
+    subprocess.run(scp_command, check=True)
+
+    # For custom image, we can use the pre-configured environment
+    setup_steps = [
+        "tar -xzf /tmp/repo.tar.gz --no-same-owner -C /workspace/",
+        f"cd /workspace/{repo_name}",
+        "source $HOME/.local/bin/env && source /opt/venv/bin/activate",
+        args.test_command
+    ]
+    
+    remote_command = " && ".join(setup_steps)
+
+    ssh_command = [
+        "ssh", "-o", "StrictHostKeyChecking=no", "-o",
+        "UserKnownHostsFile=/dev/null", "-o", "ServerAliveInterval=60", "-o",
+        "ServerAliveCountMax=10", "-p",
+        str(ssh_port), f"root@{ssh_ip}", remote_command
+    ]
+
+    print(f"Connecting to {ssh_ip}:{ssh_port}...")
+
+    try:
+        process = subprocess.Popen(ssh_command,
+                                   stdout=subprocess.PIPE,
+                                   stderr=subprocess.STDOUT,
+                                   universal_newlines=True,
+                                   bufsize=0)
+
+        stdout_lines = []
+
+        print("Command output:")
+
+        for line in iter(process.stdout.readline, ''):
+            print(line.strip())
+            stdout_lines.append(line)
+
+        process.wait()
+
+        return_code = process.returncode
+        success = return_code == 0
+
+        stdout_str = "".join(stdout_lines)
+
+        if success:
+            print("Command executed successfully")
+        else:
+            print(f"Command failed with exit code {return_code}")
+
+        result = {
+            "success": success,
+            "return_code": return_code,
+            "stdout": stdout_str,
+            "stderr": ""
+        }
+        return result
+
+    except Exception as e:
+        print(f"Error executing SSH command: {str(e)}")
+        result = {"success": False, "error": str(e), "stdout": "", "stderr": ""}
+        return result
+
+
+def terminate_pod(pod_id):
+    """Terminate the pod"""
+    print("Terminating RunPod...")
+    requests.delete(f"{PODS_API}/{pod_id}", headers=HEADERS)
+    print(f"Terminated pod {pod_id}")
+
+
+def main():
+    pod_id = None
+    try:
+        pod_id = create_pod()
+        wait_for_pod(pod_id)
+        result = execute_command(pod_id)
+
+        if result.get("error") is not None:
+            print(f"Error executing command: {result['error']}")
+            sys.exit(1)
+
+        if not result.get("success", False):
+            print(
+                "Tests failed - check the output above for details on which tests failed"
+            )
+            sys.exit(1)
+
+    finally:
+        if pod_id:
+            terminate_pod(pod_id)
+
+
+if __name__ == "__main__":
+    main()

.github/scripts/runpod_cleanup.py

@@ -0,0 +1,90 @@
+import json
+import os
+import sys
+import uuid
+
+import requests
+
+API_KEY = os.environ['RUNPOD_API_KEY']
+RUN_ID = os.environ.get('GITHUB_RUN_ID', str(uuid.uuid4()))
+PODS_API = "https://rest.runpod.io/v1/pods"
+HEADERS = {
+    "Content-Type": "application/json",
+    "Authorization": f"Bearer {API_KEY}"
+}
+
+
+def get_job_ids():
+    """Parse job IDs from environment variable"""
+    job_ids_str = os.environ.get('JOB_IDS')
+    try:
+        job_ids = json.loads(job_ids_str)
+        if not isinstance(job_ids, list):
+            print("Error: JOB_IDS is not a list.")
+            sys.exit(1)
+        return job_ids
+    except json.JSONDecodeError as e:
+        print(f"Error parsing JOB_IDS: {e}")
+        sys.exit(1)
+
+
+def cleanup_pods():
+    """Find and terminate RunPod instances"""
+    print(f"Run ID: {RUN_ID}")
+
+    single_job_id = os.environ.get('JOB_ID')
+    
+    if single_job_id:
+        job_ids = [single_job_id]
+        print(f"Job ID: {single_job_id}")
+    else:
+        job_ids = get_job_ids()
+        print(f"Job IDs: {job_ids}")
+
+    # Get all pods associated with RunPod API_KEY
+    try:
+        response = requests.get(PODS_API, headers=HEADERS)
+        response.raise_for_status()
+        pods = response.json()
+    except requests.exceptions.RequestException as e:
+        print(f"Error getting pods: {e}")
+        sys.exit(1)
+
+    # Find and terminate pods created by this workflow run
+    terminated_pods = []
+    for pod in pods:
+        pod_name = pod.get("name", "")
+        pod_id = pod.get("id")
+
+        # Check if this pod was created by one of our jobs
+        if any(f"{job_id}-{RUN_ID}" in pod_name for job_id in job_ids):
+            print(f"Found pod: {pod_id} ({pod_name})")
+            try:
+                print(f"Terminating pod {pod_id}...")
+                term_response = requests.delete(f"{PODS_API}/{pod_id}",
+                                                headers=HEADERS)
+                term_response.raise_for_status()
+                terminated_pods.append(pod_id)
+                print(f"Successfully terminated pod {pod_id}")
+            except requests.exceptions.RequestException as e:
+                print(f"Error terminating pod {pod_id}: {e}")
+                sys.exit(1)
+    
+    if terminated_pods:
+        if single_job_id:
+            print(f"Terminated pod: {terminated_pods[0]}")
+        else:
+            print(f"Terminated {len(terminated_pods)} pods: {terminated_pods}")
+    else:
+        if single_job_id:
+            print(f"No pod found matching pattern: {single_job_id}-{RUN_ID}")
+        else:
+            print("No pods found to terminate.")
+
+
+def main():
+    cleanup_pods()
+
+
+if __name__ == "__main__":
+    main()

.github/workflows/build-image.yml

@@ -32,7 +32,7 @@ permissions:
 jobs:
   build-python-3-10:
     if: ${{ github.event.inputs.python_3_10 == 'true' }}
-    uses: ./.github/workflows/_template-build-image.yml
+    uses: ./.github/workflows/build-image-template.yml
     with:
       python_version: '3.10'
       dockerfile_path: docker/Dockerfile.python3.10
@@ -41,7 +41,7 @@ jobs:
 
   build-python-3-11:
     if: ${{ github.event.inputs.python_3_11 == 'true' }}
-    uses: ./.github/workflows/_template-build-image.yml
+    uses: ./.github/workflows/build-image-template.yml
     with:
       python_version: '3.11'
       dockerfile_path: docker/Dockerfile.python3.11
@@ -50,7 +50,7 @@ jobs:
 
   build-python-3-12:
     if: ${{ github.event.inputs.python_3_12 == 'true' }}
-    uses: ./.github/workflows/_template-build-image.yml
+    uses: ./.github/workflows/build-image-template.yml
     with:
       python_version: '3.12'
       dockerfile_path: docker/Dockerfile.python3.12
@@ -59,9 +59,9 @@ jobs:
 
   build-python-3-12-cuda-12-9:
     if: ${{ github.event.inputs.python_3_12_cuda_12_9 == 'true' }}
-    uses: ./.github/workflows/_template-build-image.yml
+    uses: ./.github/workflows/build-image-template.yml
     with:
       python_version: '3.12'
       dockerfile_path: docker/Dockerfile.python3.12.cuda12.9.1
       tag_suffix: py3.12-cuda12.9.1
-    secrets: inherit
+    secrets: inherit

.github/workflows/ci-precommit.yml

@@ -1,29 +0,0 @@
-name: pre-commit
-
-on:
-  pull_request:
-    branches: [main]
-  workflow_call:
-
-concurrency:
-  group: pre-commit-${{ github.ref }}
-  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
-
-permissions:
-  contents: read
-
-jobs:
-  pre-commit:
-    if: github.event_name == 'workflow_call' || github.event.pull_request.draft != true
-    runs-on: ubuntu-latest
-    steps:
-    - uses: actions/checkout@v4
-    - uses: actions/setup-python@v5
-      with:
-        python-version: "3.12"
-    - run: echo "::add-matcher::.github/workflows/matchers/actionlint.json"
-    - run: echo "::add-matcher::.github/workflows/matchers/mypy.json"
-    - run: echo "::add-matcher::.github/workflows/matchers/ruff.json"
-    - uses: pre-commit/action@v3.0.1
-      with:
-        extra_args: --all-files --hook-stage manual

.github/workflows/ci-slash-commands.yml

@@ -1,230 +0,0 @@
-name: Slash Commands
-
-on:
-  issue_comment:
-    types: [created]
-
-permissions:
-  contents: read
-  pull-requests: write
-  statuses: write
-
-jobs:
-  handle-merge:
-    if: >-
-      github.event.issue.pull_request != null
-      && startsWith(github.event.comment.body, '/merge')
-    runs-on: ubuntu-latest
-    steps:
-      - name: Check write permission
-        id: perm
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            const { data: perm } = await github.rest.repos.getCollaboratorPermissionLevel({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              username: context.payload.comment.user.login,
-            });
-            const hasWrite = ['admin', 'write'].includes(perm.permission);
-            if (!hasWrite) {
-              core.setFailed(`User ${context.payload.comment.user.login} lacks write permission (has: ${perm.permission}).`);
-            }
-            core.setOutput('has_write', String(hasWrite));
-
-      - name: Add ready label and react
-        if: steps.perm.outputs.has_write == 'true'
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            const owner = context.repo.owner;
-            const repo = context.repo.repo;
-            const prNumber = context.payload.issue.number;
-            // Remove ready first to allow re-trigger (labeled event fires on add, not if already present)
-            try { await github.rest.issues.removeLabel({ owner, repo, issue_number: prNumber, name: 'ready' }); } catch {}
-            await github.rest.issues.addLabels({ owner, repo, issue_number: prNumber, labels: ['ready'] });
-            await github.rest.reactions.createForIssueComment({
-              owner, repo,
-              comment_id: context.payload.comment.id,
-              content: 'rocket',
-            });
-  parse-command:
-    if: >-
-      github.event.issue.pull_request != null
-      && startsWith(github.event.comment.body, '/test')
-    runs-on: ubuntu-latest
-    outputs:
-      test_type: ${{ steps.parse.outputs.test_type }}
-      test_scope: ${{ steps.parse.outputs.test_scope }}
-      full_suite: ${{ steps.parse.outputs.full_suite }}
-      pr_sha: ${{ steps.pr.outputs.sha }}
-      pr_branch: ${{ steps.pr.outputs.branch }}
-      has_write: ${{ steps.perm.outputs.has_write }}
-    steps:
-      - name: Check write permission
-        id: perm
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            const { data: perm } = await github.rest.repos.getCollaboratorPermissionLevel({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              username: context.payload.comment.user.login,
-            });
-            const hasWrite = ['admin', 'write'].includes(perm.permission);
-            core.setOutput('has_write', String(hasWrite));
-            if (!hasWrite) {
-              core.info(`User ${context.payload.comment.user.login} lacks write permission — ignoring.`);
-            }
-
-      - name: Parse /test command
-        id: parse
-        if: steps.perm.outputs.has_write == 'true'
-        shell: bash
-        env:
-          COMMENT: ${{ github.event.comment.body }}
-        run: |
-          set -euo pipefail
-          TEST_NAME=$(echo "$COMMENT" | grep -oP '(?<=/test\s)\S+' | head -1 || true)
-
-          VALID="encoder vae transformer kernel unit ssim training lora-inference lora-training distillation self-forcing vsa vmoba performance api full fastcheck pre-commit"
-          if [ -z "$TEST_NAME" ] || ! echo "$VALID" | grep -qw "$TEST_NAME"; then
-            echo "Unknown test: '$TEST_NAME'. Valid: $VALID"
-            exit 1
-          fi
-
-          declare -A MAP=(
-            [encoder]=encoder [vae]=vae [transformer]=transformer
-            [kernel]=kernel_tests [unit]=unit_test
-            [ssim]=ssim [training]=training
-            [lora-inference]=inference_lora [lora-training]=training_lora
-            [distillation]=distillation_dmd [self-forcing]=self_forcing
-            [vsa]=training_vsa [vmoba]=inference_vmoba
-            [performance]=performance [api]=api_server
-          )
-
-          if [ "$TEST_NAME" = "full" ]; then
-            {
-              echo "test_type=all"
-              echo "test_scope=full"
-              echo "full_suite=true"
-            } >> "$GITHUB_OUTPUT"
-          elif [ "$TEST_NAME" = "fastcheck" ]; then
-            {
-              echo "test_type=fastcheck"
-              echo "test_scope=fastcheck"
-              echo "full_suite=false"
-            } >> "$GITHUB_OUTPUT"
-          elif [ "$TEST_NAME" = "pre-commit" ]; then
-            {
-              echo "test_type="
-              echo "test_scope=precommit"
-              echo "full_suite=false"
-            } >> "$GITHUB_OUTPUT"
-          else
-            {
-              echo "test_type=${MAP[$TEST_NAME]}"
-              echo "test_scope=direct"
-              echo "full_suite=false"
-            } >> "$GITHUB_OUTPUT"
-          fi
-
-      - name: Get PR details
-        id: pr
-        if: steps.perm.outputs.has_write == 'true'
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            const { data: pr } = await github.rest.pulls.get({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              pull_number: context.payload.issue.number,
-            });
-            core.setOutput('sha', pr.head.sha);
-            core.setOutput('branch', pr.head.ref);
-
-  pre-commit:
-    needs: parse-command
-    if: >-
-      needs.parse-command.outputs.has_write == 'true'
-      && needs.parse-command.outputs.test_scope == 'precommit'
-    uses: ./.github/workflows/ci-precommit.yml
-
-  post-precommit-status:
-    needs: [parse-command, pre-commit]
-    if: always() && needs.parse-command.outputs.test_scope == 'precommit'
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        env:
-          PR_SHA: ${{ needs.parse-command.outputs.pr_sha }}
-          RESULT: ${{ needs.pre-commit.result }}
-        with:
-          script: |
-            const state = process.env.RESULT === 'success' ? 'success' : 'failure';
-            await github.rest.repos.createCommitStatus({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              sha: process.env.PR_SHA,
-              state,
-              context: 'pre-commit',
-              description: `Triggered via /test pre-commit (${state})`,
-            });
-
-  trigger-buildkite:
-    needs: parse-command
-    if: >-
-      needs.parse-command.outputs.has_write == 'true'
-      && needs.parse-command.outputs.test_type != ''
-    runs-on: ubuntu-latest
-    steps:
-      - name: React to comment
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            await github.rest.reactions.createForIssueComment({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              comment_id: context.payload.comment.id,
-              content: 'rocket',
-            });
-
-      - name: Trigger Buildkite
-        env:
-          BUILDKITE_API_TOKEN: ${{ secrets.BUILDKITE_API_TOKEN }}
-          PR_SHA: ${{ needs.parse-command.outputs.pr_sha }}
-          PR_BRANCH: ${{ needs.parse-command.outputs.pr_branch }}
-          PR_NUMBER: ${{ github.event.issue.number }}
-          TEST_SCOPE: ${{ needs.parse-command.outputs.test_scope }}
-          FULL_SUITE: ${{ needs.parse-command.outputs.full_suite }}
-          TEST_TYPE: ${{ needs.parse-command.outputs.test_type }}
-          BK_ORG: ${{ vars.BUILDKITE_ORG_SLUG }}
-          BK_PIPELINE: ${{ vars.BUILDKITE_PIPELINE_SLUG }}
-        run: |
-          curl -sS --fail-with-body -X POST \
-            "https://api.buildkite.com/v2/organizations/${BK_ORG}/pipelines/${BK_PIPELINE}/builds" \
-            -H "Authorization: Bearer $BUILDKITE_API_TOKEN" \
-            -H "Content-Type: application/json" \
-            --data-raw "$(jq -n \
-              --arg commit "$PR_SHA" \
-              --arg branch "$PR_BRANCH" \
-              --arg message "/test ${TEST_TYPE} on PR #${PR_NUMBER}" \
-              --argjson pr_id "$PR_NUMBER" \
-              --arg test_scope "$TEST_SCOPE" \
-              --arg full_suite "$FULL_SUITE" \
-              --arg test_type "$TEST_TYPE" \
-              --arg pr_number "$PR_NUMBER" \
-              '{
-                commit: $commit,
-                branch: $branch,
-                message: $message,
-                ignore_pipeline_branch_filters: true,
-                pull_request_id: $pr_id,
-                pull_request_base_branch: "main",
-                env: {
-                  TEST_SCOPE: $test_scope,
-                  FULL_SUITE: $full_suite,
-                  TEST_TYPE: $test_type,
-                  PR_NUMBER: $pr_number
-                }
-              }')"

.github/workflows/ci-trigger-full-suite.yml

@@ -1,83 +0,0 @@
-name: Trigger Full Suite
-
-on:
-  pull_request:
-    types: [labeled, synchronize]
-
-permissions:
-  contents: read
-  pull-requests: read
-
-concurrency:
-  group: full-suite-${{ github.event.pull_request.number }}
-  cancel-in-progress: true
-
-jobs:
-  trigger:
-    if: >-
-      (github.event.action == 'labeled' && github.event.label.name == 'ready')
-      || github.event.action == 'synchronize'
-    runs-on: ubuntu-latest
-    steps:
-      - name: Check ready label
-        id: check
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            const { data: pr } = await github.rest.pulls.get({
-              owner: context.repo.owner,
-              repo: context.repo.repo,
-              pull_number: context.payload.pull_request.number,
-            });
-            const hasReady = pr.labels.some(l => l.name === 'ready');
-            core.setOutput('has_ready', String(hasReady));
-            if (!hasReady) core.info('No ready label — skipping Full Suite trigger.');
-
-      - name: Cancel previous Buildkite builds
-        if: steps.check.outputs.has_ready == 'true'
-        env:
-          BUILDKITE_API_TOKEN: ${{ secrets.BUILDKITE_API_TOKEN }}
-          PR_BRANCH: ${{ github.event.pull_request.head.ref }}
-        run: |
-          # Find running builds for this branch with TEST_SCOPE=full and cancel them
-          builds=$(curl -sS -H "Authorization: Bearer $BUILDKITE_API_TOKEN" \
-            "https://api.buildkite.com/v2/organizations/${{ vars.BUILDKITE_ORG_SLUG }}/pipelines/${{ vars.BUILDKITE_PIPELINE_SLUG }}/builds?branch=${PR_BRANCH}&state=running,scheduled" \
-            | jq -r '.[] | select(.env.TEST_SCOPE == "full") | .number')
-          for build_num in $builds; do
-            echo "Cancelling Buildkite build #$build_num"
-            curl -sS -X PUT -H "Authorization: Bearer $BUILDKITE_API_TOKEN" \
-              "https://api.buildkite.com/v2/organizations/${{ vars.BUILDKITE_ORG_SLUG }}/pipelines/${{ vars.BUILDKITE_PIPELINE_SLUG }}/builds/${build_num}/cancel"
-          done
-
-      - name: Trigger Buildkite Full Suite
-        if: steps.check.outputs.has_ready == 'true'
-        env:
-          BUILDKITE_API_TOKEN: ${{ secrets.BUILDKITE_API_TOKEN }}
-          PR_SHA: ${{ github.event.pull_request.head.sha }}
-          PR_BRANCH: ${{ github.event.pull_request.head.ref }}
-          PR_NUMBER: ${{ github.event.pull_request.number }}
-          BK_ORG: ${{ vars.BUILDKITE_ORG_SLUG }}
-          BK_PIPELINE: ${{ vars.BUILDKITE_PIPELINE_SLUG }}
-        run: |
-          curl -sS --fail-with-body -X POST \
-            "https://api.buildkite.com/v2/organizations/${BK_ORG}/pipelines/${BK_PIPELINE}/builds" \
-            -H "Authorization: Bearer $BUILDKITE_API_TOKEN" \
-            -H "Content-Type: application/json" \
-            --data-raw "$(jq -n \
-              --arg commit "$PR_SHA" \
-              --arg branch "$PR_BRANCH" \
-              --arg message "Full Suite for PR #${PR_NUMBER}" \
-              --argjson pr_id "$PR_NUMBER" \
-              '{
-                commit: $commit,
-                branch: $branch,
-                message: $message,
-                ignore_pipeline_branch_filters: true,
-                pull_request_id: $pr_id,
-                pull_request_base_branch: "main",
-                env: {
-                  TEST_SCOPE: "full",
-                  FULL_SUITE: "true",
-                  PR_NUMBER: ($pr_id | tostring)
-                }
-              }')"

.github/workflows/community-issue-labeler.yml

@@ -1,65 +0,0 @@
-name: Auto-Label Issues
-
-on:
-  issues:
-    types: [opened, edited]
-
-permissions:
-  issues: write
-
-jobs:
-  label-issues:
-    if: github.repository == 'hao-ai-lab/FastVideo'
-    runs-on: ubuntu-latest
-    steps:
-      - name: Label by keywords
-        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
-        with:
-          script: |
-            const title = context.payload.issue.title.toLowerCase();
-            const body = (context.payload.issue.body || '').toLowerCase();
-            const text = title + ' ' + body;
-            const labels = [];
-
-            const rules = [
-              // scope labels (shared with PR labeling via Mergify)
-              // Mapping: label → repo directories
-              //   scope: training    → fastvideo/train/, fastvideo/training/, fastvideo/distillation/
-              //   scope: inference   → fastvideo/pipelines/, fastvideo/entrypoints/, fastvideo/worker/
-              //   scope: attention   → fastvideo/attention/
-              //   scope: kernel      → fastvideo-kernel/, csrc/
-              //   scope: model       → fastvideo/models/, fastvideo/layers/, fastvideo/configs/models/
-              //   scope: data        → fastvideo/dataset/, fastvideo/pipelines/preprocess/
-              //   scope: distributed → fastvideo/distributed/
-              //   scope: docs        → docs/
-              { keywords: ['training', 'finetune', 'fine-tune', 'lora', 'fsdp', 'distill'], label: 'scope: training' },
-              { keywords: ['inference', 'generate', 'pipeline', 'slow', 'latency'], label: 'scope: inference' },
-              { keywords: ['attention', 'vsa', 'flash', 'sta', 'vmoba', 'sparse attn'], label: 'scope: attention' },
-              { keywords: ['kernel', 'csrc', 'cuda kernel', 'thunderkittens'], label: 'scope: kernel' },
-              { keywords: ['wan', 'hunyuan', 'mochi', 'ltx', 'cogvideo', 'flux', 'sd3', 'cosmos'], label: 'scope: model' },
-              { keywords: ['dataset', 'dataloader', 'preprocessing', 'preprocess'], label: 'scope: data' },
-              { keywords: ['distributed', 'sequence parallel', 'fsdp', 'tensor parallel', 'multi-node', 'multi-gpu'], label: 'scope: distributed' },
-              { keywords: ['docs', 'documentation', 'tutorial', 'example'], label: 'scope: docs' },
-              // issue-only labels (cross-module, no single repo directory)
-              { keywords: ['install', 'setup', 'pip', 'cuda', 'uv ', 'import error', 'modulenotfound'], label: 'installation' },
-              { keywords: ['memory', 'oom', 'out of memory', 'gpu memory', 'vram'], label: 'performance' },
-              { keywords: ['windows', 'macos', 'mac os', 'apple', 'mps', 'rocm', 'amd', 'npu'], label: 'platform' },
-            ];
-
-            for (const rule of rules) {
-              if (rule.keywords.some(kw => text.includes(kw))) {
-                labels.push(rule.label);
-              }
-            }
-
-            if (labels.length > 0) {
-              await github.rest.issues.addLabels({
-                owner: context.repo.owner,
-                repo: context.repo.repo,
-                issue_number: context.payload.issue.number,
-                labels: labels,
-              });
-              console.log(`Added labels: ${labels.join(', ')}`);
-            } else {
-              console.log('No keyword matches found');
-            }

.github/workflows/community-stale.yml

@@ -1,51 +0,0 @@
-name: Close Stale Issues and PRs
-
-on:
-  schedule:
-    # Daily at 1:30 AM UTC
-    - cron: '30 1 * * *'
-
-jobs:
-  stale:
-    if: github.repository == 'hao-ai-lab/FastVideo'
-    permissions:
-      issues: write
-      pull-requests: write
-      actions: write
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/stale@997185467fa4f803885201cee163a9f38240193d # v10.1.1
-        with:
-          operations-per-run: 500
-
-          exempt-draft-pr: true
-          exempt-issue-labels: 'keep-open,pinned,security,Bug,RFC'
-          exempt-pr-labels: 'keep-open,pinned'
-
-          labels-to-add-when-unstale: 'unstale'
-          labels-to-remove-when-stale: 'unstale'
-
-          days-before-issue-stale: 90
-          days-before-issue-close: 30
-          stale-issue-label: 'stale'
-          stale-issue-message: >
-            This issue has been automatically marked as stale because it has not
-            had any activity within 90 days. It will be automatically closed if
-            no further activity occurs within 30 days. Leave a comment if you
-            feel this issue should remain open. Thank you!
-          close-issue-message: >
-            This issue has been automatically closed due to inactivity. Please
-            feel free to reopen if you feel it is still relevant. Thank you!
-
-          days-before-pr-stale: 60
-          days-before-pr-close: 14
-          stale-pr-label: 'stale'
-          stale-pr-message: >
-            This pull request has been automatically marked as stale because it
-            has not had any activity within 60 days. It will be automatically
-            closed if no further activity occurs within 14 days. Leave a comment
-            if you feel this pull request should remain open. Thank you!
-          close-pr-message: >
-            This pull request has been automatically closed due to inactivity.
-            Please feel free to reopen if you intend to continue working on it.
-            Thank you!

.github/workflows/community-welcome.yml

@@ -1,56 +0,0 @@
-name: Welcome First-Time Contributors
-
-on:
-  issues:
-    types: [opened]
-  pull_request_target:
-    types: [opened]
-
-permissions:
-  issues: write
-  pull-requests: write
-
-jobs:
-  welcome:
-    if: github.repository == 'hao-ai-lab/FastVideo'
-    runs-on: ubuntu-latest
-    steps:
-      - uses: actions/first-interaction@34f15e814fe48ac9312ccf29db4e74fa767cbab7 # v1.3.0
-        with:
-          repo-token: ${{ secrets.GITHUB_TOKEN }}
-          issue-message: |
-            Welcome to FastVideo! Thanks for opening your first issue.
-
-            To help us investigate, please include:
-            - **FastVideo version**: `pip show fastvideo`
-            - **GPU**: `nvidia-smi` output (GPU model, driver, CUDA version)
-            - **Python version**: `python --version`
-            - **OS**: e.g., Ubuntu 22.04
-
-            If this is a bug, a minimal reproduction script helps us fix it faster.
-
-            Useful links:
-            - [Documentation](https://hao-ai-lab.github.io/FastVideo)
-            - [Contributing Guide](https://hao-ai-lab.github.io/FastVideo/contributing/overview/)
-            - [Slack](https://join.slack.com/t/fastvideo/shared_invite/zt-3f4lao1uq-u~Ipx6Lt4J27AlD2y~IdLQ)
-          pr-message: |
-            Welcome to FastVideo! Thanks for your first pull request.
-
-            **How our CI works:**
-
-            PRs run a two-tier CI system:
-            1. **Pre-commit** — formatting (yapf), linting (ruff), type checking (mypy). Runs immediately on every PR.
-            2. **Fastcheck** — core GPU tests (encoders, VAEs, transformers, kernels, unit tests). Runs automatically via Buildkite on relevant file changes (~10-15 min).
-            3. **Full Suite** — integration tests, training pipelines, SSIM regression. Runs only when a reviewer adds the `ready` label.
-
-            **Before your PR is reviewed:**
-            - [ ] `pre-commit run --all-files` passes locally
-            - [ ] You've added or updated tests for your changes
-            - [ ] The PR description explains what and why
-
-            If pre-commit fails, a bot comment will explain how to fix it. Fastcheck and Full Suite results appear in the Checks section below.
-
-            **Useful links:**
-            - [Contributing Guide](https://hao-ai-lab.github.io/FastVideo/contributing/overview/)
-            - [Development Roadmap](https://github.com/hao-ai-lab/FastVideo/issues/899)
-            - [Slack](https://join.slack.com/t/fastvideo/shared_invite/zt-3f4lao1uq-u~Ipx6Lt4J27AlD2y~IdLQ)

.github/workflows/docs.yml

@@ -7,14 +7,14 @@ on:
       - 'docs/**'
       - 'mkdocs.yml'
       - 'requirements-mkdocs.txt'
-      - '.github/workflows/infra-docs.yml'
+      - '.github/workflows/docs.yml'
   pull_request:
     branches: [ main ]
     paths:
       - 'docs/**'
       - 'mkdocs.yml'
       - 'requirements-mkdocs.txt'
-      - '.github/workflows/infra-docs.yml'
+      - '.github/workflows/docs.yml'
 
 permissions:
   contents: read

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

@@ -36,11 +36,11 @@ jobs:
           
           if [ "$NEW_VERSION" != "$OLD_VERSION" ]; then
             echo "Version changed from $OLD_VERSION to $NEW_VERSION"
-            echo "changed=true" >> "$GITHUB_OUTPUT"
-            echo "new-version=$NEW_VERSION" >> "$GITHUB_OUTPUT"
+            echo "changed=true" >> $GITHUB_OUTPUT
+            echo "new-version=$NEW_VERSION" >> $GITHUB_OUTPUT
           else
             echo "Version did not change"
-            echo "changed=false" >> "$GITHUB_OUTPUT"
+            echo "changed=false" >> $GITHUB_OUTPUT
           fi
 
   build_wheels:

.github/workflows/matchers/ruff.json

@@ -1,17 +0,0 @@
-{
-  "problemMatcher": [
-    {
-      "owner": "ruff",
-      "pattern": [
-        {
-          "regexp": "^(.+):(\\d+):(\\d+): (\\w+) (.+)$",
-          "file": 1,
-          "line": 2,
-          "column": 3,
-          "code": 4,
-          "message": 5
-        }
-      ]
-    }
-  ]
-}

.github/workflows/pr-test.yml

@@ -0,0 +1,338 @@
+name: PR Test
+
+on:
+  push:
+    branches: [main]
+    paths:
+      - "fastvideo/**/*.py"
+      - ".github/workflows/pr-test.yml"
+  pull_request:
+    branches: [main]
+    types: [opened, ready_for_review, synchronize, reopened]
+    paths:
+      - "fastvideo/**/*.py"
+      - ".github/workflows/pr-test.yml"
+      - "pyproject.toml"
+      - "docker/Dockerfile.python3.12"
+      - "csrc/**"
+  workflow_dispatch:
+    inputs:
+      run_encoder_test:
+        description: "Run encoder-test"
+        required: false
+        default: false
+        type: boolean
+      run_vae_test:
+        description: "Run vae-test"
+        required: false
+        default: false
+        type: boolean
+      run_transformer_test:
+        description: "Run transformer-test"
+        required: false
+        default: false
+        type: boolean
+      run_ssim_test:
+        description: "Run ssim-test"
+        required: false
+        default: false
+        type: boolean
+      run_training_test:
+        description: "Run training-test"
+        required: false
+        default: false
+        type: boolean
+      run_training_test_VSA:
+        description: "Run training-test-VSA"
+        required: false
+        default: false
+        type: boolean
+      run_precision_test_VSA:
+        description: "Run precision-test-VSA"
+        required: false
+        default: false
+        type: boolean
+      run_unit_test:
+        description: "Run unit-test"
+        required: false
+        default: false
+        type: boolean
+
+env:
+  PYTHONUNBUFFERED: "1"
+
+
+concurrency:
+  group: pr-test-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  pre-commit:
+    uses: ./.github/workflows/pre-commit.yml
+
+  change-filter:
+    runs-on: ubuntu-latest
+    needs: pre-commit
+    if: ${{ github.event.pull_request.draft == false || github.event_name == 'workflow_dispatch' }}
+    outputs:
+      encoder-test: ${{ steps.filter.outputs.encoder-test }}
+      vae-test: ${{ steps.filter.outputs.vae-test }}
+      transformer-test: ${{ steps.filter.outputs.transformer-test }}
+      training-test: ${{ steps.filter.outputs.training-test }}
+      training-test-VSA: ${{ steps.filter.outputs.training-test-VSA }}
+      precision-test-VSA: ${{ steps.filter.outputs.precision-test-VSA }}
+      unit-test: ${{ steps.filter.outputs.unit-test }}
+    steps:
+      - uses: actions/checkout@v4
+      - uses: dorny/paths-filter@v3
+        id: filter
+        with:
+          filters: |
+            # Define reusable path patterns
+            common-paths: &common-paths
+              - 'pyproject.toml'
+              - 'docker/Dockerfile.python3.10'
+              - 'docker/Dockerfile.python3.11'
+              - 'docker/Dockerfile.python3.12'
+            vsa-kernel-paths: &vsa-kernel-paths
+              - 'csrc/attn/video_sparse_attn/**'
+              - 'csrc/attn/video_sparse_attn/tk/**'
+              - 'csrc/attn/video_sparse_attn/setup.py'
+              - 'csrc/attn/video_sparse_attn/config_vsa.py'
+              - 'csrc/attn/video_sparse_attn/vsa.cpp'
+            vsa-paths: &vsa-paths
+              - 'fastvideo/**'
+              - *common-paths
+              - *vsa-kernel-paths
+            
+            # Actual tests
+            encoder-test:
+              - 'fastvideo/models/encoders/**'
+              - 'fastvideo/models/loader/**'
+              - 'fastvideo/tests/encoders/**'
+              - *common-paths
+            vae-test:
+              - 'fastvideo/models/vaes/**'
+              - 'fastvideo/models/loader/**'
+              - 'fastvideo/tests/vaes/**'
+              - *common-paths
+            transformer-test:
+              - 'fastvideo/models/dits/**'
+              - 'fastvideo/models/loader/**'
+              - 'fastvideo/tests/transformers/**'
+              - 'fastvideo/layers/**'
+              - 'fastvideo/attention/**'
+              - *common-paths
+            training-test:
+              - 'fastvideo/**'
+              - *common-paths
+            training-test-VSA:
+              - 'fastvideo/**'
+              - *common-paths
+              - *vsa-kernel-paths
+            precision-test-VSA:
+              - *common-paths
+              - *vsa-kernel-paths
+            unit-test:
+              - 'fastvideo/**'
+              - *common-paths
+
+  encoder-test:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.encoder-test == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_encoder_test == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "encoder-test"
+      gpu_type: "NVIDIA A40"
+      gpu_count: 1
+      volume_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "uv pip install -e .[test] && pytest ./fastvideo/tests/encoders -s"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+
+  vae-test:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.vae-test == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_vae_test == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "vae-test"
+      gpu_type: "NVIDIA A40"
+      gpu_count: 1
+      volume_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "uv pip install -e .[test] && pytest ./fastvideo/tests/vaes -s"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+
+  transformer-test:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.transformer-test == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_transformer_test == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "transformer-test"
+      gpu_type: "NVIDIA L40S"
+      gpu_count: 1
+      volume_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "uv pip install -e .[test] && pytest ./fastvideo/tests/transformers -s"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+      
+  ssim-test:
+    needs: change-filter
+    if: >-
+      github.event_name != 'workflow_dispatch' || (github.event_name == 'workflow_dispatch' && github.event.inputs.run_ssim_test == 'true')
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: [
+          # {version: "3.10", tag: "latest"},
+          # {version: "3.11", tag: "py3.11-latest"},
+          {version: "3.12", tag: "py3.12-latest"}
+        ]
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "ssim-test-py${{ matrix.python-version.version }}"
+      gpu_type: "NVIDIA A40"
+      gpu_count: 2
+      volume_size: 200
+      disk_size: 200
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:${{ matrix.python-version.tag }}"
+      test_command: "uv pip install -e .[test] && pytest ./fastvideo/tests/ssim -vs"
+      timeout_minutes: 60
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+
+  training-test:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.training-test == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_training_test == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "training-test"
+      gpu_type: "NVIDIA A40"
+      gpu_count: 4
+      volume_size: 100
+      disk_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "wandb login $WANDB_API_KEY && uv pip install -e .[test] && pytest ./fastvideo/tests/training/Vanilla -srP"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+      WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
+
+  training-test-VSA:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.training-test-VSA == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_training_test_VSA == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "training-test-VSA"
+      gpu_type: "NVIDIA H100 NVL"
+      gpu_count: 2
+      volume_size: 100
+      disk_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "wandb login $WANDB_API_KEY && uv pip install -e .[test] && pytest ./fastvideo/tests/training/VSA -srP"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+      WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
+
+  precision-test-VSA:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.precision-test-VSA == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_precision_test_VSA == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "precision-test-VSA"
+      gpu_type: "NVIDIA H100 NVL"
+      gpu_count: 1
+      volume_size: 100
+      disk_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "uv pip install -e .[test] && python csrc/attn/tests/test_vsa.py"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+
+  unit-test:
+    needs: change-filter
+    if: >-
+      (github.event_name != 'workflow_dispatch' && needs.change-filter.outputs.unit-test == 'true') || 
+      (github.event_name == 'workflow_dispatch' && github.event.inputs.run_unit_test == 'true')
+    uses: ./.github/workflows/runpod-test.yml
+    with:
+      job_id: "unit-test"
+      gpu_type: "NVIDIA L40S"
+      gpu_count: 1
+      volume_size: 100
+      disk_size: 100
+      image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+      test_command: "uv pip install -e .[test] && pytest ./fastvideo/dataset/ -vs && pytest ./fastvideo/workflow/ -vs && pytest ./fastvideo/entrypoints/ -vs"
+      timeout_minutes: 30
+    secrets:
+      RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+      RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+
+  # nightly-test:
+  #   if: >-
+  #     (github.event_name == 'workflow_dispatch' && github.event.inputs.run_nightly_test == 'true')
+  #   uses: ./.github/workflows/runpod-test.yml
+  #   with:
+  #     job_id: "nightly-test"
+  #     gpu_type: "NVIDIA A40"
+  #     gpu_count: 4
+  #     volume_size: 100
+  #     disk_size: 100
+  #     image: "ghcr.io/${{ github.repository }}/fastvideo-dev:py3.12-latest"
+  #     test_command: "wandb login $WANDB_API_KEY && uv pip install -e .[test] && pytest ./fastvideo/tests/nightly/test_e2e_overfit_single_sample.py -vs"
+  #     timeout_minutes: 30
+  #   secrets:
+  #     RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+  #     RUNPOD_PRIVATE_KEY: ${{ secrets.RUNPOD_PRIVATE_KEY }}
+  #     WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
+
+  runpod-cleanup:
+    # Add other jobs to this list as you create them
+    needs: [encoder-test, vae-test, transformer-test, ssim-test, training-test, training-test-VSA, precision-test-VSA] 
+    if: ${{ always() && ((github.event_name != 'workflow_dispatch' && github.event.pull_request.draft == false) || github.event_name == 'workflow_dispatch') }}
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.10"
+
+      - name: Install dependencies
+        run: pip install requests
+
+      - name: Cleanup all RunPod instances
+        env:
+          JOB_IDS: '["encoder-test", "vae-test", "transformer-test", "ssim-test-py3.10", "ssim-test-py3.11", "ssim-test-py3.12", "training-test", "training-test-VSA", "precision-test-VSA"]'
+          RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+          GITHUB_RUN_ID: ${{ github.run_id }}
+        run: python .github/scripts/runpod_cleanup.py

.github/workflows/pre-commit.yml

@@ -0,0 +1,18 @@
+name: pre-commit
+
+on:
+    workflow_call:
+
+jobs:
+  pre-commit:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v4
+    - uses: actions/setup-python@v5
+      with:
+        python-version: "3.12"
+    - run: echo "::add-matcher::.github/workflows/matchers/actionlint.json"
+    - run: echo "::add-matcher::.github/workflows/matchers/mypy.json"
+    - uses: pre-commit/action@v3.0.1
+      with:
+        extra_args: --all-files --hook-stage manual

.github/workflows/runpod-test.yml

@@ -0,0 +1,94 @@
+name: RunPod Test
+
+on:
+  workflow_call:
+    inputs:
+      job_id:
+        required: true
+        type: string
+        description: "Unique identifier for this test job"
+      gpu_type:
+        required: true
+        type: string
+        description: "GPU type to use (e.g. NVIDIA A40, NVIDIA L40S)"
+      gpu_count:
+        required: true
+        type: number
+        description: "Number of GPUs to use"
+      volume_size:
+        required: false
+        type: number
+        default: 20
+        description: "Volume size in GB"
+      disk_size:
+        required: false
+        type: number
+        default: 20
+        description: "Disk size in GB"
+      image:
+        required: true
+        type: string
+        description: "Docker image to use"
+      test_command:
+        required: true
+        type: string
+        description: "Command to run tests"
+      timeout_minutes:
+        required: false
+        type: number
+        default: 30
+        description: "Timeout in minutes"
+    secrets:
+      RUNPOD_API_KEY:
+        required: true
+      RUNPOD_PRIVATE_KEY:
+        required: true
+      WANDB_API_KEY:
+        required: false
+
+jobs:
+  run-test:
+    runs-on: ubuntu-latest
+    environment: runpod-runners
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Set up SSH key
+        run: |
+          mkdir -p ~/.ssh
+          echo "${{ secrets.RUNPOD_PRIVATE_KEY }}" > ~/.ssh/id_rsa
+          chmod 600 ~/.ssh/id_rsa
+          ssh-keygen -y -f ~/.ssh/id_rsa > ~/.ssh/id_rsa.pub
+
+      - name: Install dependencies
+        run: pip install requests
+
+      - name: Run tests on RunPod
+        env:
+          JOB_ID: ${{ inputs.job_id }}
+          RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+          GITHUB_RUN_ID: ${{ github.run_id }}
+          WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
+        timeout-minutes: ${{ inputs.timeout_minutes }}
+        run: >-
+          python .github/scripts/runpod_api.py 
+          --gpu-type "${{ inputs.gpu_type }}"
+          --gpu-count ${{ inputs.gpu_count }}
+          --volume-size ${{ inputs.volume_size }}
+          --disk-size ${{ inputs.disk_size }}
+          --image "${{ inputs.image }}"
+          --test-command "${{ inputs.test_command }}"
+
+      - name: Terminate RunPod Instances
+        if: ${{ always() }}
+        env:
+          RUNPOD_API_KEY: ${{ secrets.RUNPOD_API_KEY }}
+          GITHUB_RUN_ID: ${{ github.run_id }}
+          JOB_ID: ${{ inputs.job_id }}
+        run: python .github/scripts/runpod_cleanup.py 

.github/workflows/test.yml

@@ -0,0 +1,31 @@
+name: Run Tests
+
+on:
+  push:
+    branches: [ main ]
+  pull_request:
+    branches: [ main ]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check out repository
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'  # or any version you need
+
+      - name: Install dependencies
+        run: |
+          python -m pip install --upgrade pip setuptools wheel
+          pip install torch
+          pip install packaging ninja
+          pip install -e .
+          pip install pytest
+
+      - name: Run Pytest
+        run: |
+          pytest --ignore csrc/attn/test

.github/workflows/vsa-publish.yml

@@ -0,0 +1,257 @@
+name: Publish Video Sparse Attention Kernel to PyPI on Version Change
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - "csrc/attn/video_sparse_attn/setup.py"
+  workflow_dispatch:
+
+jobs:
+  check-version-change:
+    runs-on: ubuntu-latest
+    outputs:
+      version-changed: ${{ steps.check-version.outputs.changed }}
+      new-version: ${{ steps.check-version.outputs.new-version }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 2
+
+      - name: Check if version changed
+        id: check-version
+        run: |
+          cd csrc/attn/video_sparse_attn
+          # Get current commit's version
+          NEW_VERSION=$(grep -oP 'VERSION\s*=\s*"\K[^"]+' setup.py)
+          echo "New version: $NEW_VERSION"
+          
+          # Get previous version from git history
+          OLD_VERSION=$(git show HEAD~1:./setup.py | grep -oP 'VERSION\s*=\s*"\K[^"]+' || echo "0.0.0")
+          echo "Old version: $OLD_VERSION"
+          
+          if [ "$NEW_VERSION" != "$OLD_VERSION" ]; then
+            echo "Version changed from $OLD_VERSION to $NEW_VERSION"
+            echo "changed=true" >> $GITHUB_OUTPUT
+            echo "new-version=$NEW_VERSION" >> $GITHUB_OUTPUT
+          else
+            echo "Version did not change"
+            echo "changed=false" >> $GITHUB_OUTPUT
+          fi
+
+  build_wheels:
+    name: Build Wheel
+    needs: check-version-change
+    if: ${{ needs.check-version-change.outputs.version-changed == 'true' || github.event_name == 'workflow_dispatch' }}
+    runs-on: ${{ matrix.os }}
+
+    strategy:
+      fail-fast: false
+      matrix:
+          # Using ubuntu-20.04 instead of 22.04 for more compatibility (glibc). Ideally we'd use the
+          # manylinux docker image, but I haven't figured out how to install CUDA on manylinux.
+          os: [ubuntu-22.04]
+          python-version: ['3.10', '3.11', '3.12', '3.13']
+          # For version reference https://pytorch.org/get-started/previous-versions/
+          torch-cuda:
+            - torch-version: '2.5.1'
+              cuda-version: '12.4.1'
+              torch-cuda-short: 'cu124'
+            - torch-version: '2.6.0'
+              cuda-version: '12.6.3'
+              torch-cuda-short: 'cu126'
+            - torch-version: '2.7.1'
+              cuda-version: '12.8.0'
+              torch-cuda-short: 'cu128'
+
+    steps:
+      - name: Free up disk space
+        run: |
+          echo "Initial disk space:"
+          df -h
+          
+          # Remove large directories
+          sudo rm -rf /usr/share/dotnet
+          sudo rm -rf /usr/local/lib/android
+          sudo rm -rf /opt/ghc
+          sudo rm -rf /usr/local/share/boost
+          sudo rm -rf /usr/share/swift
+          sudo rm -rf /usr/local/lib/node_modules
+          sudo rm -rf /usr/local/share/powershell
+          sudo rm -rf /usr/share/rust
+          sudo rm -rf /usr/local/.ghcup
+          
+          # Remove cached files
+          sudo rm -rf /var/lib/apt/lists/*
+          sudo rm -rf /var/cache/apt/archives/*
+          
+          echo "Disk space after cleanup:"
+          df -h
+
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Set up Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install CUDA ${{ matrix.torch-cuda.cuda-version }}
+        uses: Jimver/cuda-toolkit@v0.2.21
+        id: cuda-toolkit
+        with:
+          cuda: ${{ matrix.torch-cuda.cuda-version }}
+          linux-local-args: '["--toolkit"]'
+          method: 'network'
+
+      - name: Install dependencies (GCC, Clang, CUDA Paths, Git)
+        run: |
+          sudo apt update
+          sudo apt install -y git patchelf gcc-11 g++-11 clang-11
+          sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-11 100 --slave /usr/bin/g++ g++ /usr/bin/g++-11
+
+          # Allow Git to Access Safe Directory
+          git config --global --add safe.directory /__w/FastVideo/FastVideo
+
+          # Set CUDA environment variables
+          export CUDA_HOME=/usr/local/cuda-${{ matrix.torch-cuda.cuda-version }}
+          export PATH=${CUDA_HOME}/bin:${PATH}
+          export LD_LIBRARY_PATH=${CUDA_HOME}/lib64:$LD_LIBRARY_PATH
+          
+          # Verify installation
+          gcc --version
+          g++ --version
+          clang-11 --version
+          nvcc --version
+
+      - name: Install PyTorch ${{ matrix.torch-cuda.torch-version }}+cu${{ matrix.torch-cuda.cuda-version }}
+        run: |
+          pip install --upgrade pip
+          # With python 3.13 and torch 2.5.1, unless we update typing-extensions, we get error
+          # AttributeError: attribute '__default__' of 'typing.ParamSpec' objects is not writable
+          pip install typing-extensions==4.12.2
+          # We want to figure out the CUDA version to download pytorch
+          # e.g. we can have system CUDA version being 11.7 but if torch==1.12 then we need to download the wheel from cu116
+          # see https://github.com/pytorch/pytorch/blob/main/RELEASE.md#release-compatibility-matrix
+          pip install --no-cache-dir torch==${{ matrix.torch-cuda.torch-version }} --index-url https://download.pytorch.org/whl/${{matrix.torch-cuda.torch-cuda-short}}
+          nvcc --version
+          python --version
+          python -c "import torch; print('PyTorch:', torch.__version__)"
+          python -c "import torch; print('CUDA:', torch.version.cuda)"
+          python -c "from torch.utils import cpp_extension; print (cpp_extension.CUDA_HOME)"
+
+      - name: Build wheel
+        run: |
+          export PYTHONPATH=$GITHUB_WORKSPACE:$PYTHONPATH
+          
+          # We want setuptools >= 49.6.0 otherwise we can't compile the extension if system CUDA version is 11.7 and pytorch cuda version is 11.6
+          # https://github.com/pytorch/pytorch/blob/664058fa83f1d8eede5d66418abff6e20bd76ca8/torch/utils/cpp_extension.py#L810
+          # However this still fails so I'm using a newer version of setuptools
+          pip install setuptools
+          pip install ninja packaging wheel
+
+          cd csrc/attn/video_sparse_attn  # Move into the correct folder
+          git submodule update --init --recursive # Ensure ThunderKittens submodule is initialized
+          python setup.py bdist_wheel --dist-dir=dist
+
+      - name: Rename wheel file
+        run: |
+          cd csrc/attn/video_sparse_attn
+
+          CUDA_SHORT_VERSION=$(echo ${{ matrix.torch-cuda.cuda-version }} | cut -d. -f1,2 | sed 's/\.//g')
+          TORCH_SHORT_VERSION=$(echo ${{ matrix.torch-cuda.torch-version }} | cut -d. -f1,2)
+          # Get the correct version format
+          tmpname=cu${CUDA_SHORT_VERSION}torch${TORCH_SHORT_VERSION}
+          wheel_name=$(ls dist/*whl | xargs -n 1 basename | sed "s/-/+$tmpname-/2")
+          # Rename with version information
+          ls dist/*whl |xargs -I {} mv {} dist/${wheel_name}
+          echo "wheel_name=${wheel_name}" >> $GITHUB_ENV
+
+      - name: Upload wheel artifact
+        uses: actions/upload-artifact@v4
+        with:
+          name: ${{ env.wheel_name }}
+          path: csrc/attn/video_sparse_attn/dist/*.whl
+          retention-days: 90
+
+  publish_package:
+    name: Publish package
+    needs: [build_wheels, check-version-change]
+    if: ${{ needs.check-version-change.outputs.version-changed == 'true' || github.event_name == 'workflow_dispatch' }}
+    runs-on: ubuntu-22.04
+    permissions:
+      id-token: write  # Needed for OIDC Trusted Publishing
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: '3.10'
+
+      - name: Install CUDA 12.4.1
+        uses: Jimver/cuda-toolkit@v0.2.21
+        id: cuda-toolkit
+        with:
+          cuda: 12.4.1
+          linux-local-args: '["--toolkit"]'
+          method: 'network'
+          sub-packages: '["nvcc"]'
+
+      - name: Install dependencies (GCC, Clang, CUDA Paths, Git)
+        run: |
+          sudo apt update
+          sudo apt install -y git patchelf gcc-11 g++-11 clang-11
+          sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-11 100 --slave /usr/bin/g++ g++ /usr/bin/g++-11
+
+          # Allow Git to Access Safe Directory
+          git config --global --add safe.directory /__w/FastVideo/FastVideo
+
+          # Set CUDA environment variables
+          export CUDA_HOME=/usr/local/cuda-12.4.1
+          export PATH=${CUDA_HOME}/bin:${PATH}
+          export LD_LIBRARY_PATH=${CUDA_HOME}/lib64:$LD_LIBRARY_PATH
+          
+          # Verify installation
+          gcc --version
+          g++ --version
+          clang-11 --version
+          nvcc --version
+
+      - name: Install PyTorch 2.5.1+cu12.4.1
+        run: |
+          pip install --upgrade pip
+          # With python 3.13 and torch 2.5.1, unless we update typing-extensions, we get error
+          # AttributeError: attribute '__default__' of 'typing.ParamSpec' objects is not writable
+          pip install typing-extensions==4.12.2
+          # We want to figure out the CUDA version to download pytorch
+          # e.g. we can have system CUDA version being 11.7 but if torch==1.12 then we need to download the wheel from cu116
+          # see https://github.com/pytorch/pytorch/blob/main/RELEASE.md#release-compatibility-matrix
+          export TORCH_CUDA_VERSION=124
+          pip install --no-cache-dir torch==2.5.1 --index-url https://download.pytorch.org/whl/cu${TORCH_CUDA_VERSION}
+          nvcc --version
+          python --version
+          python -c "import torch; print('PyTorch:', torch.__version__)"
+          python -c "import torch; print('CUDA:', torch.version.cuda)"
+          python -c "from torch.utils import cpp_extension; print (cpp_extension.CUDA_HOME)"
+
+      - name: Build source distribution
+        run: |
+          export PYTHONPATH=$GITHUB_WORKSPACE:$PYTHONPATH
+          
+          # We want setuptools >= 49.6.0 otherwise we can't compile the extension if system CUDA version is 11.7 and pytorch cuda version is 11.6
+          # https://github.com/pytorch/pytorch/blob/664058fa83f1d8eede5d66418abff6e20bd76ca8/torch/utils/cpp_extension.py#L810
+          # However this still fails so I'm using a newer version of setuptools
+          pip install setuptools
+          pip install ninja packaging wheel
+
+          cd csrc/attn/video_sparse_attn  # Move into the correct folder
+          git submodule update --init --recursive # Ensure ThunderKittens submodule is initialized
+          python setup.py sdist --dist-dir=dist
+
+      - name: Publish release distributions to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: csrc/attn/video_sparse_attn/dist/

.gitignore

@@ -85,9 +85,6 @@ docs/distillation/examples/
 dmd_t2v_output/
 preprocess_output_text/
 
-
 .claude/
 .codex/
-.sisyphus/
 openspec/
-fastvideo/tests/ssim/reference_videos/**

.pre-commit-config.yaml

@@ -18,8 +18,10 @@ exclude: |
     fastvideo/train\.py|
     fastvideo/utils/.*|
     examples/.*|
-    .github/workflows/publish-fastvideo.yml|
-    .github/workflows/_template-build-image.yml|
+    .github/workflows/fastvideo-publish.yml|
+    .github/workflows/sta-publish.yml|
+    .github/workflows/vsa-publish.yml|
+    .github/workflows/build-image-template.yml|
     docs/source/inference/support_matrix.md
   )
 repos:

README.md

@@ -9,7 +9,7 @@
 **FastVideo is a unified post-training and real-time inference framework for accelerated video generation.**
 
 ## NEWS
-- `2026/03/17`: Release Live demo: [Into the Dreamverse: Vibe Directing in FastVideo](https://dreamverse.fastvideo.org/), check out the [Blog](https://haoailab.com/blogs/dreamverse/).
+
 - `2026/03/13`: Release Live demo: [Create a 5s 1080p Video in 4.5s with FastVideo on a Single GPU](https://1080p.fastvideo.org/), check out the [Blog](https://haoailab.com/blogs/fastvideo_realtime_1080p/).
 - `2025/11/19`: Release [CausalWan2.2 I2V A14B Preview](https://huggingface.co/FastVideo/CausalWan2.2-I2V-A14B-Preview-Diffusers) models, [Blog](https://hao-ai-lab.github.io/blogs/fastvideo_causalwan_preview/) and [Inference Code!](https://github.com/hao-ai-lab/FastVideo/blob/main/examples/inference/basic/basic_self_forcing_causal_wan2_2_i2v.py).
 - `2025/08/04`: Release [FastWan](https://hao-ai-lab.github.io/FastVideo/distillation/dmd) models and [Sparse-Distillation](https://hao-ai-lab.github.io/blogs/fastvideo_post_training/).

docs/README.md

@@ -36,4 +36,4 @@ Then open your browser to: http://localhost:8000
 
 ## Automatic Deployment
 
-Documentation is automatically built and deployed to GitHub Pages when changes are pushed to the `main` branch via the `.github/workflows/infra-docs.yml` workflow.
+Documentation is automatically built and deployed to GitHub Pages when changes are pushed to the `main` branch via the `.github/workflows/docs.yml` workflow.

docs/contributing/ci_architecture.md

@@ -1,306 +0,0 @@
-# CI Architecture
-
-## Overview
-
-FastVideo uses a three-tier CI pipeline designed to keep feedback fast on every push while
-protecting `main` through a full GPU regression suite before any merge.
-
-```
-PR push
-  │
-  ├─► Tier 1: Pre-commit (~2 min)
-  │     GitHub Actions / ubuntu-latest
-  │     yapf, ruff, mypy, codespell, pymarkdown, actionlint, check-filenames
-  │
-  └─► Tier 2: Fastcheck (~10-20 min, path-filtered)
-        Buildkite / Modal GPU instances
-        Only runs tests for paths you changed
-
-              │ (developer comments /merge or maintainer adds 'ready' label)
-              ▼
-        Tier 3: Full Suite (~60-90 min)
-          Buildkite / Modal GPU instances
-          All integration, SSIM, training, and performance tests
-          Runs on the PR branch directly
-              │
-          pass ──► Mergify auto-squash-merges to main, branch deleted
-          fail ──► Mergify removes 'ready' label; fix and /merge again
-```
-
----
-
-## CI Tiers
-
-### Tier 1: Pre-commit (every PR push)
-
-| Attribute | Value |
-|-----------|-------|
-| Triggered by | Every push to any PR branch, plus pushes to `main` |
-| Runs on | GitHub Actions, `ubuntu-latest` |
-| Duration | ~2 minutes |
-
-**Checks run** (from `.pre-commit-config.yaml`, stage: `manual`):
-
-| Hook | What it checks |
-|------|---------------|
-| `yapf` | Python code formatting |
-| `ruff` | Python linting and auto-fixable style issues |
-| `codespell` | Spelling errors in code and docs |
-| `pymarkdown` | Markdown formatting |
-| `actionlint` | GitHub Actions workflow syntax |
-| `mypy` | Static type checking (Python 3.10 target) |
-| `check-filenames` | No spaces in tracked filenames |
-
-A failure here means code style or type issues. Run `pre-commit run --all-files` locally to
-replicate CI results before pushing.
-
----
-
-### Tier 2: Fastcheck (path-filtered, every PR push)
-
-| Attribute | Value |
-|-----------|-------|
-| Triggered by | Every push; the monorepo-diff plugin skips tests for unchanged paths |
-| Runs on | Buildkite, Modal GPU instances |
-| Duration | ~10-20 minutes per test (run in parallel) |
-
-**Tests and their path triggers:**
-
-| Buildkite label | `TEST_TYPE` | Triggers when you change |
-|-----------------|-------------|--------------------------|
-| Encoder Tests | `encoder` | `fastvideo/models/encoders/**`, `fastvideo/models/loader/**`, `fastvideo/tests/encoders/**`, `pyproject.toml`, `docker/Dockerfile.python3.12` |
-| VAE Tests | `vae` | `fastvideo/models/vaes/**`, `fastvideo/models/loader/**`, `fastvideo/tests/vaes/**`, `pyproject.toml`, `docker/Dockerfile.python3.12` |
-| Transformer Tests | `transformer` | `fastvideo/models/dits/**`, `fastvideo/models/loader/**`, `fastvideo/tests/transformers/**`, `fastvideo/layers/**`, `fastvideo/attention/**`, `pyproject.toml`, `docker/Dockerfile.python3.12` |
-| Kernel Tests | `kernel_tests` | `fastvideo-kernel/**`, `pyproject.toml`, `docker/Dockerfile.python3.12` |
-| Unit Tests | `unit_test` | `fastvideo/**`, `.buildkite/**`, `.github/**`, `pyproject.toml`, `docker/Dockerfile.python3.12` |
-
-A Fastcheck failure means a component-level regression. Check the Buildkite build log for the
-failing test's output.
-
----
-
-### Tier 3: Full Test Suite (triggered by `ready` label)
-
-| Attribute | Value |
-|-----------|-------|
-| Triggered by | Adding the `ready` label to the PR (via `/merge` command), or a `/test full` command |
-| Runs on | Buildkite, Modal GPU instances |
-| Duration | 60-90 minutes total (tests run in parallel, path-filtered) |
-
-**Tests:**
-
-| Buildkite label | `TEST_TYPE` | Timeout |
-|-----------------|-------------|---------|
-| SSIM Tests | `ssim` | 90 min |
-| LoRA Inference Tests | `inference_lora` | 20 min |
-| Training Tests | `training` | 15 min |
-| Distillation DMD Tests | `distillation_dmd` | 15 min |
-| Self-Forcing Tests | `self_forcing` | 15 min |
-| LoRA Training Tests | `training_lora` | 15 min |
-| Training Tests VSA | `training_vsa` | 15 min |
-| Inference Tests VMoBA | `inference_vmoba` | 15 min |
-| Performance Tests | `performance` | 30 min |
-| API Server Tests | `api_server` | 30 min |
-
-A Full Suite failure removes the `ready` label automatically. A Mergify comment links to
-the Buildkite build. Fix the regression, push, and comment `/merge` again.
-
----
-
-## Auto-merge Flow
-
-Mergify prevents untested code from landing on `main` by gating squash-merge on the Full
-Suite passing directly on the PR branch.
-
-**How it works:**
-
-1. A developer comments `/merge` on an approved PR (or a maintainer adds the `ready` label).
-2. The `ready` label triggers `ci-trigger-full-suite.yml`, which calls the Buildkite API to
-   run the Full Suite on the PR branch itself.
-3. While the Full Suite runs, Mergify also auto-rebases the PR branch against `main` if it
-   is behind and has no conflicts.
-4. Once the Full Suite posts `full-suite-passed`, Mergify checks all **merge conditions**:
-   - `pre-commit` check is green
-   - `fastcheck-passed` check is green
-   - `full-suite-passed` check is green
-   - At least 1 approved review (`#approved-reviews-by>=1`)
-   - PR title starts with a valid `[type]` tag
-   - PR is not a draft
-   - No merge conflicts
-5. If all conditions pass, Mergify squash-merges to `main` automatically. The branch is
-   deleted after merge.
-6. If the Full Suite fails, Mergify removes the `ready` label and posts a comment linking to
-   the Buildkite build. The developer fixes the issue, pushes, and comments `/merge` again.
-
-**Merge conditions summary:**
-
-| Condition | Meaning |
-|-----------|---------|
-| `check-success~=pre-commit` | Tier 1 pre-commit must be green |
-| `check-success=fastcheck-passed` | Tier 2 Fastcheck must be green |
-| `check-success=full-suite-passed` | Tier 3 Full Suite must be green |
-| `#approved-reviews-by>=1` | At least one approved review |
-| `title~=(?i)^\[(feat|bugfix|...)` | PR title has a valid type tag |
-| `-draft` | PR is not in draft state |
-| `-conflict` | No merge conflicts with base branch |
-| `-closed` | PR is still open |
-
----
-
-## Label System
-
-Labels are applied automatically. You don't need to set them manually.
-
-### Type Labels (from PR title prefix)
-
-Applied by Mergify based on the `[tag]` at the start of the PR title.
-
-| Label | Matched title prefix | Meaning |
-|-------|---------------------|---------|
-| `type: feat` | `[feat]` or `[feature]` | New feature or capability |
-| `type: bugfix` | `[bugfix]` or `[fix]` | Bug fix |
-| `type: refactor` | `[refactor]` | Code restructuring, no behavior change |
-| `type: perf` | `[perf]` | Performance improvement |
-| `type: ci` | `[ci]` | CI/CD or tooling changes |
-| `type: docs` | `[doc]` or `[docs]` | Documentation only |
-| `type: misc` | `[misc]` or `[chore]` | Housekeeping, dependency bumps |
-| `type: new-model` | `[new-model]` | Adding a new model |
-
-### Scope Labels (from changed files)
-
-Applied by Mergify based on which paths you modified. Multiple scope labels can be added.
-
-| Label | File paths that trigger it |
-|-------|---------------------------|
-| `scope: training` | `fastvideo/train/`, `fastvideo/training/`, `fastvideo/distillation/`, `examples/train/`, `examples/training/`, `examples/distill/` |
-| `scope: inference` | `fastvideo/pipelines/basic/`, `fastvideo/pipelines/stages/`, `fastvideo/pipelines/samplers/`, `fastvideo/entrypoints/`, `fastvideo/worker/`, `fastvideo/configs/sample/`, `fastvideo/configs/pipelines/`, `examples/inference/` |
-| `scope: attention` | `fastvideo/attention/` |
-| `scope: kernel` | `fastvideo-kernel/`, `csrc/` |
-| `scope: data` | `fastvideo/dataset/`, `fastvideo/pipelines/preprocess/`, `examples/preprocessing/` |
-| `scope: infra` | `.github/`, `.buildkite/`, `fastvideo/tests/`, `docker/` |
-| `scope: distributed` | `fastvideo/distributed/` |
-| `scope: docs` | `docs/` |
-| `scope: ui` | `ui/` |
-| `scope: model` | `fastvideo/models/`, `fastvideo/layers/`, `fastvideo/configs/models/` |
-
-### Process Labels
-
-| Label | Who sets it | Meaning |
-|-------|-------------|---------|
-| `ready` | Developer (`/merge` command) or maintainer | Triggers Full Suite and enables auto-merge |
-| `needs-rebase` | Mergify (automatic) | PR has merge conflicts; rebase needed |
-| `do-not-merge` | Maintainer | Blocks queue entry regardless of other conditions |
-
----
-
-## PR Title Format
-
-All PR titles targeting `main` must start with a bracketed type tag. This is enforced by a
-Mergify merge protection rule and is required before a PR can be squash-merged.
-
-**Format:**
-
-```
-[type] Short description
-```
-
-**Valid type tags:**
-
-`feat`, `feature`, `bugfix`, `fix`, `refactor`, `perf`, `ci`, `doc`, `docs`, `misc`, `chore`,
-`kernel`, `new-model`
-
-**Valid examples:**
-
-```
-[feat] Add causal Wan pipeline
-[bugfix] Fix VAE temporal tiling corruption
-[refactor] Restructure training framework
-[perf] Optimize FlashAttention kernel dispatch
-[docs] Add inference guide for LoRA
-[new-model] Port HunyuanVideo 1.5 to FastVideo
-```
-
-**Invalid examples (will block merge):**
-
-```
-Add causal Wan pipeline          ← missing type tag
-FEAT: Add pipeline               ← wrong format
-feat: Add pipeline               ← square brackets required
-```
-
-If your title is invalid, Mergify posts a comment explaining the required format and the
-merge protection check will remain failed until you update the title.
-
----
-
-## Slash Commands
-
-Slash commands let contributors and maintainers trigger CI actions directly from PR comments.
-**Write permission to the repository is required.**
-
-The command is recognized within a few seconds. The workflow reacts with a 🚀 emoji to confirm.
-
-### `/merge`
-
-```
-/merge
-```
-
-Adds the `ready` label to the PR, which triggers the Full Suite on your PR branch and
-enables Mergify to auto-squash-merge once all conditions pass.
-
-The command first removes the `ready` label if it is already present, then re-adds it. This
-ensures the `labeled` event fires and a fresh Full Suite build is started even on a re-try.
-
-### `/test <name>`
-
-```
-/test <name>
-```
-
-Triggers a specific Buildkite test or suite on the current PR branch.
-
-| Command | Runs | Maps to `TEST_TYPE` |
-|---------|------|---------------------|
-| `/test encoder` | Encoder Tests (Fastcheck) | `encoder` |
-| `/test vae` | VAE Tests (Fastcheck) | `vae` |
-| `/test transformer` | Transformer Tests (Fastcheck) | `transformer` |
-| `/test kernel` | Kernel Tests (Fastcheck) | `kernel_tests` |
-| `/test unit` | Unit Tests (Fastcheck) | `unit_test` |
-| `/test ssim` | SSIM regression tests | `ssim` |
-| `/test training` | Training pipeline tests | `training` |
-| `/test lora-inference` | LoRA inference tests | `inference_lora` |
-| `/test lora-training` | LoRA training tests | `training_lora` |
-| `/test distillation` | DMD distillation tests | `distillation_dmd` |
-| `/test self-forcing` | Self-Forcing tests | `self_forcing` |
-| `/test vsa` | VSA training tests | `training_vsa` |
-| `/test vmoba` | VMoBA inference tests | `inference_vmoba` |
-| `/test performance` | Performance benchmarks | `performance` |
-| `/test api` | API server integration tests | `api_server` |
-| `/test full` | Entire Full Suite | all (with `TEST_SCOPE=full`) |
-| `/test fastcheck` | Entire Fastcheck suite | fastcheck (with `TEST_SCOPE=fastcheck`) |
-
----
-
-## Auto Branch Cleanup
-
-After a PR is squash-merged to `main`, Mergify automatically deletes the head branch.
-Protected branches (`main`, `master`, `release/*`) are never deleted.
-
----
-
-## Workflow File Reference
-
-| Filename | Trigger | What it does |
-|----------|---------|-------------|
-| `ci-precommit.yml` | Every push / PR against `main` | Runs pre-commit hooks (yapf, ruff, mypy, codespell, pymarkdown, actionlint, check-filenames) |
-| `ci-trigger-full-suite.yml` | `ready` label added to a PR | Calls Buildkite API to run Full Suite on the PR branch |
-| `ci-slash-commands.yml` | PR comment starting with `/merge` or `/test` | Handles slash commands; adds `ready` label or triggers Buildkite |
-| `community-issue-labeler.yml` | Issue opened or edited | Auto-labels issues by keyword matching against title and body |
-| `community-welcome.yml` | First contribution | Posts a welcome comment for first-time contributors |
-| `community-stale.yml` | Scheduled | Marks and closes stale issues and PRs |
-| `infra-build-image.yml` | Manual (`workflow_dispatch`) | Builds Docker images for CI |
-| `infra-docs.yml` | Changes to `docs/` merged to `main` | Builds and deploys documentation to GitHub Pages |
-| `publish-fastvideo.yml` | Version bump | Publishes `fastvideo` package to PyPI |
-| `publish-kernel.yml` | Version bump | Publishes `fastvideo-kernel` package to PyPI |
-| `publish-comfyui.yml` | Version bump | Publishes ComfyUI node package to PyPI |

docs/contributing/developer_env/runpod.md

@@ -1,195 +1,65 @@
-# RunPod Development Environment
 
-RunPod gives you on-demand cloud GPUs for FastVideo development. It's useful when you need a beefy GPU to test training runs, benchmark inference, or reproduce results without waiting for shared cluster time.
+# 📦 Developing FastVideo on RunPod
 
-## Prerequisites
+You can easily use the FastVideo Pod Template on [RunPod](https://www.runpod.io) for development or experimentation.
 
-- A [RunPod](https://www.runpod.io) account with billing configured
-- An SSH key pair. If you don't have one, generate it with `ssh-keygen -t ed25519`
-- Your public key (`~/.ssh/id_ed25519.pub`) ready to paste into RunPod
-
-## Step 1: Create a Pod
-
-**1. Verify your account**
-
-Make sure you're logged into the right RunPod account before spending credits.
+## Creating a new pod
 
+- Make sure you are using the correct RunPod account.
 ![RunPod Account Selection](../../assets/images/runpod_account.png)
 
-**2. Filter by CUDA version**
-
-Use "Additional Filters" to select CUDA 12.8.
-
+- Use "Additional Filters" to select CUDA 12.8.
 ![RunPod CUDA selection](../../assets/images/runpod_cuda.png)
 
-**3. Select a GPU**
-
-Click "Deploy" and pick a GPU. See [GPU Recommendations](#gpu-recommendations) below for guidance on which GPU to choose.
-
+- Click "Deploy" and Pick a single A40 or RTX 4090 GPU.
 ![RunPod GPU Selection](../../assets/images/runpod_deploy.png)
 
-**4. Pick the FastVideo template**
-
-Select the "FastVideo" or "fastvideo-dev" Pod Template. This pulls the pre-built image that includes all dependencies, Flash Attention, and a ready-to-use `uv` environment.
-
+- Select the "FastVideo" or "fastvideo-dev" Pod Template.
 ![RunPod Pod Template Selection](../../assets/images/runpod_create.png)
 
-**5. Name your pod**
-
-Use a memorable name like `yourname-fastvideo-2026-03-28`. This helps if you have multiple pods running.
-
-**6. Add a persistent volume (recommended)**
-
-Attach a network volume to `/root/.cache` or `/models` for storing downloaded model weights. Models can be 10-50 GB each, and re-downloading them every session wastes time and bandwidth.
-
-**7. Deploy**
-
-Click Deploy. The pod takes a few minutes to start while the image pulls. You'll see it transition to "Running" in your dashboard.
-
-## Step 2: Connect via SSH
-
-Once the pod is running, find the "SSH over exposed TCP" connection string in the pod dashboard.
+- Set the Pod name to "`<name>-<FastVideo>-<date>`".
 
+- Finally, once the pod is deployed (will take a few minutes as the image is being pulled), you can SSH into it using "SSH exposed over TCP". You'll need to use the matching private ssh key you provided.
 ![RunPod SSH](../../assets/images/runpod_ssh.png)
 
-Connect with:
+## Working with the pod
+
+After SSH'ing into your pod, you'll find the correct `uv` environment already activated and you should be in /FastVideo/ directory. Make sure to use /FastVideo/ for all your work.
+
+To pull in the latest changes from the GitHub repo:
 
 ```bash
-ssh root@<pod-ip> -p <port> -i ~/.ssh/id_ed25519
+cd /FastVideo
+git pull
 ```
 
-RunPod also supports VS Code Remote SSH if you prefer an IDE.
+Run your development workflows as usual:
 
-### Custom template (advanced)
+```bash
+# Run linters
+pre-commit run --all-files
 
-If you're setting up a pod from scratch instead of the FastVideo template, use this image:
+# Run tests
+pytest tests/
+```
+
+Make sure to push your changes back to the GitHub repo as nothing will be saved to the pod when it is terminated.
+
+After you are done with your work, you can terminate the pod by clicking the "Terminate" and "Delete" buttons. Remember if the pod is not completely deleted, Runpod will keep charging you for it.
+
+## Extra Information:
+If you need to customize the pod template this section has some useful information. For the most part you can leave the defaults of the FastVideo Pod Template.
+
+When creating your pod template, use this image:
 
 ```
 ghcr.io/hao-ai-lab/fastvideo/fastvideo-dev:py3.12-latest
 ```
 
-And paste this as the Container Start Command to enable SSH ([RunPod docs](https://docs.runpod.io/pods/configuration/use-ssh)):
+Paste Container Start Command to support SSH ([RunPod Docs](https://docs.runpod.io/pods/configuration/use-ssh)):
 
 ```bash
 bash -c "apt update;DEBIAN_FRONTEND=noninteractive apt-get install openssh-server -y;mkdir -p ~/.ssh;cd $_;chmod 700 ~/.ssh;echo \"$PUBLIC_KEY\" >> authorized_keys;chmod 700 authorized_keys;service ssh start;sleep infinity"
 ```
 
 ![RunPod template configuration](../../assets/images/runpod_template.png)
-
-## Step 3: Set Up FastVideo
-
-After SSH'ing in, the `uv` virtual environment at `/opt/venv` is already activated (configured in `.bashrc` and `.profile`). You land in the `/FastVideo` directory.
-
-**Clone or pull the repo**
-
-If the pod already has the FastVideo repo:
-
-```bash
-cd /FastVideo
-git pull
-```
-
-If starting from a blank pod:
-
-```bash
-git clone https://github.com/hao-ai-lab/FastVideo.git /FastVideo
-cd /FastVideo
-```
-
-**Install the package**
-
-```bash
-uv pip install -e .[dev]
-```
-
-The Docker image already includes Flash Attention and most heavy dependencies, so this is fast.
-
-**Build the custom kernels (optional)**
-
-VSA and STA attention kernels aren't in the Docker image by default. Build them if you're working on attention backends or need maximum inference performance:
-
-```bash
-cd /FastVideo/fastvideo-kernel
-./build.sh
-```
-
-The build script detects your GPU architecture automatically. An A100 or H100 takes about 5-10 minutes.
-
-**Verify the setup**
-
-```bash
-cd /FastVideo
-python -c "import fastvideo; print('OK')"
-pytest tests/ -q --no-header
-```
-
-## Development Workflow
-
-### Editing code on RunPod
-
-Two common approaches:
-
-**Option A: Edit on RunPod directly**
-
-Use VS Code Remote SSH or `vim`/`nano` on the pod. Commit and push when you're ready:
-
-```bash
-cd /FastVideo
-git add .
-git commit -m "your change"
-git push
-```
-
-**Option B: Edit locally, sync to RunPod**
-
-Work in your local repo, then pull on the pod:
-
-```bash
-# On RunPod:
-cd /FastVideo
-git pull
-```
-
-This keeps your local tools (editor, linters) intact while running GPU workloads on the pod.
-
-### Running linters and tests
-
-```bash
-# Lint
-pre-commit run --all-files
-
-# Full test suite
-pytest tests/
-
-# Just package tests
-pytest fastvideo/tests/ -v
-```
-
-### Storing models
-
-If you attached a persistent volume, point your model downloads there:
-
-```bash
-export HF_HOME=/models/huggingface
-export TRANSFORMERS_CACHE=/models/huggingface
-```
-
-Add these to `/root/.bashrc` so they persist across SSH sessions. The volume survives pod termination, so you only download models once.
-
-### Terminating the pod
-
-When you're done, push any commits you want to keep. RunPod does not save pod storage after termination.
-
-Go to your RunPod dashboard, click "Terminate", then "Delete". A pod that's stopped but not deleted still charges you for storage. Fully delete it to stop all charges.
-
-## GPU Recommendations
-
-| GPU | VRAM | Good for |
-|-----|------|----------|
-| RTX 4090 | 24 GB | Inference testing, small model fine-tuning, quick iteration |
-| A40 | 48 GB | Mid-size training runs, 480p video generation |
-| A100 (40 GB) | 40 GB | Multi-GPU inference, training with sequence parallelism |
-| A100 (80 GB) | 80 GB | Large model training, 720p+ video generation |
-| H100 | 80 GB | Heavy training, benchmarking, kernel development |
-
-For most development work, a single RTX 4090 or A40 is sufficient and cost-effective. Use A100/H100 when you need to reproduce training results at scale or test multi-GPU features.

docs/contributing/pull_requests.md

@@ -1,218 +0,0 @@
-# Contributing via Pull Requests
-
-This guide walks through the PR workflow: title format, labels, CI pipeline, and getting
-your changes merged.
-
----
-
-## PR Title Format (Required)
-
-Every PR targeting `main` must start with a type tag in square brackets. This is checked by
-Mergify before any merge is allowed.
-
-**Format:**
-
-```
-[type] Short description of the change
-```
-
-**Valid type tags:**
-
-| Tag | When to use |
-|-----|-------------|
-| `[feat]` or `[feature]` | New feature or capability |
-| `[bugfix]` or `[fix]` | Bug fix |
-| `[refactor]` | Code restructuring with no behavior change |
-| `[perf]` | Performance improvement |
-| `[ci]` | CI/CD or build tooling changes |
-| `[doc]` or `[docs]` | Documentation only |
-| `[misc]` or `[chore]` | Housekeeping, dependency bumps, minor cleanup |
-| `[kernel]` | CUDA kernel changes in `fastvideo-kernel/` |
-| `[new-model]` | Adding a new model or pipeline |
-
-**Examples:**
-
-```
-[feat] Add causal Wan 2.2 I2V pipeline
-[bugfix] Fix VAE temporal tiling corruption on H100
-[refactor] Restructure distributed attention dispatch
-[docs] Add LoRA finetuning guide
-[new-model] Port HunyuanVideo 1.5 to FastVideo
-```
-
-If your title is missing the tag, Mergify will post a comment listing the valid formats.
-Update the title and the check will re-evaluate automatically.
-
----
-
-## Labels
-
-Labels are applied automatically based on your PR title and the files you changed. You don't
-need to set them manually.
-
-**Type label** — set from the `[tag]` in your title:
-`type: feat`, `type: bugfix`, `type: refactor`, `type: perf`, `type: ci`, `type: docs`,
-`type: misc`, `type: new-model`
-
-**Scope labels** — set from which files you modified (multiple labels can apply):
-`scope: training`, `scope: inference`, `scope: attention`, `scope: kernel`, `scope: data`,
-`scope: infra`, `scope: distributed`, `scope: docs`, `scope: ui`, `scope: model`
-
-**Process labels** — set during review and merge:
-
-| Label | Who sets it | Meaning |
-|-------|-------------|---------|
-| `ready` | You (`/merge` comment) or a maintainer | Triggers Full Suite and enables auto-merge |
-| `needs-rebase` | Mergify (automatic) | Your PR has conflicts; rebase against `main` |
-| `do-not-merge` | Maintainer | Blocks merge regardless of CI status |
-
----
-
-## CI Pipeline
-
-Three tiers run automatically on every PR.
-
-**Tier 1: Pre-commit (~2 min) — runs on every push**
-
-GitHub Actions checks formatting, linting, type correctness, and spelling using pre-commit
-hooks: yapf, ruff, mypy, codespell, pymarkdown, actionlint, and check-filenames.
-
-**Tier 2: Fastcheck (~10-20 min) — runs on every push, path-filtered**
-
-Buildkite runs GPU tests only for the components you changed. If you only modified
-`fastvideo/models/vaes/`, only VAE Tests run. Tests run in parallel.
-
-**Tier 3: Full Suite (~60-90 min) — triggered by the `ready` label**
-
-When you comment `/merge` (or a maintainer adds the `ready` label), Buildkite runs the
-complete test suite on your PR branch: SSIM regression, LoRA inference and training,
-distillation, self-forcing, VSA, VMoBA, performance benchmarks, and API server tests.
-
----
-
-## Getting Your PR Merged
-
-**Step-by-step:**
-
-1. Open a PR with a title that starts with a valid `[type]` tag.
-2. Push your changes. Pre-commit and Fastcheck run automatically.
-3. Fix any pre-commit failures locally (`pre-commit run --all-files`) and push again.
-4. Wait for at least one approving review.
-5. Once approved and pre-commit is green, comment `/merge` on the PR.
-6. The `ready` label is added, which triggers the Full Suite on your PR branch.
-7. Mergify also auto-rebases your branch against `main` if it is behind and conflict-free.
-8. If all Full Suite tests pass and all merge conditions are met (approval, valid title,
-   pre-commit green, fastcheck green, no draft, no conflicts), Mergify squash-merges to
-   `main` automatically. Your branch is deleted.
-9. If a Full Suite test fails, Mergify removes the `ready` label and posts a comment with a
-   link to the Buildkite build. Fix the issue, push, and comment `/merge` again.
-
-!!! note
-    Only contributors with write permission to the repository can trigger slash commands.
-    If you're an external contributor, ask a maintainer to run `/merge` or add the `ready`
-    label for you.
-
----
-
-## Running Tests On Demand
-
-Comment on your PR to trigger specific tests independently of the auto-merge flow.
-
-**Trigger the entire Full Suite:**
-
-```
-/test full
-```
-
-**Trigger the Fastcheck suite:**
-
-```
-/test fastcheck
-```
-
-**Trigger individual tests:**
-
-```
-/test encoder          # Encoder component tests
-/test vae              # VAE component tests
-/test transformer      # Transformer / DiT tests
-/test kernel           # CUDA kernel tests
-/test unit             # Unit tests
-
-/test ssim             # SSIM regression tests
-/test training         # Training pipeline tests
-/test lora-inference   # LoRA inference tests
-/test lora-training    # LoRA training tests
-/test distillation     # DMD distillation tests
-/test self-forcing     # Self-Forcing distillation tests
-/test vsa              # VSA training tests
-/test vmoba            # VMoBA inference tests
-/test performance      # Performance benchmarks
-/test api              # API server integration tests
-```
-
-The workflow reacts with a 🚀 emoji to confirm the command was received.
-
----
-
-## Troubleshooting
-
-### Pre-commit fails
-
-Run locally to reproduce and auto-fix:
-
-```bash
-# Install pre-commit if needed
-uv pip install pre-commit
-pre-commit install
-
-# Run all checks on all files
-pre-commit run --all-files
-```
-
-Common quick fixes:
-
-- **yapf**: `yapf -i <file>` (Python formatting)
-- **ruff**: `ruff check --fix <file>` (linting)
-- **codespell**: `codespell --write-changes <file>` (spelling)
-
-### PR title format check fails
-
-Update your title to start with a valid type tag. The Mergify merge protection check
-re-evaluates automatically after you save the title.
-
-Valid tags: `feat`, `feature`, `bugfix`, `fix`, `refactor`, `perf`, `ci`, `doc`, `docs`,
-`misc`, `chore`, `kernel`, `new-model`
-
-### My PR has merge conflicts (`needs-rebase` label)
-
-Rebase against `main` and force-push:
-
-```bash
-git fetch origin main
-git rebase origin/main
-# Resolve any conflicts, then:
-git push --force-with-lease
-```
-
-Mergify removes the `needs-rebase` label automatically once conflicts are resolved.
-
-### Full Suite failed after `/merge`
-
-The Full Suite found a regression. Mergify removes the `ready` label and posts a comment
-linking to the Buildkite build. Check the failing step's output for assertion errors or
-tracebacks.
-
-Common causes:
-
-- Test failures caused by your code changes
-- Missing dependency in `pyproject.toml`
-- GPU memory issue (some tests require specific hardware like L40S or H100)
-- Kernel build failure (if you changed `fastvideo-kernel/`)
-
-After fixing, push and comment `/merge` again.
-
-### I'm an external contributor without write permission
-
-You can't use slash commands directly. After your PR is approved, ask a maintainer to
-comment `/merge` or add the `ready` label.

docs/contributing/testing.md

@@ -23,18 +23,13 @@ SSIM tests are located in `fastvideo/tests/ssim`. These tests generate videos us
 
 ```
 fastvideo/tests/ssim/
-├── reference_videos/
-│   ├── default/
-│   │   └── <GPU>_reference_videos/
-│   │       ├── <Model_Name>/
-│   │       │   ├── <Backend>/        # e.g., FLASH_ATTN, TORCH_SDPA
-│   │       │   │   └── <Video_File>
-│   └── full_quality/
-│       └── <GPU>_reference_videos/
+├── <GPU>_reference_videos/   # Reference videos organized by GPU type (e.g., L40S_reference_videos)
+│   ├── <Model_Name>/
+│   │   ├── <Backend>/        # e.g., FLASH_ATTN, TORCH_SDPA
+│   │   │   └── <Video_File>
 ├── test_causal_similarity.py
-├── test_wan_t2v_similarity.py
-├── test_wan_i2v_similarity.py
-├── reference_videos_cli.py
+├── test_inference_similarity.py
+├── update_reference_videos.sh
 └── ...
 ```
 
@@ -42,7 +37,7 @@ fastvideo/tests/ssim/
 
 To add a new SSIM test, follow these steps:
 
-1. **Create or Update a Test File**: Prefer model-specific files (for example `test_wan_t2v_similarity.py`) and create a new one when testing a distinct model or pipeline.
+1. **Create or Update a Test File**: You can add a new test function to an existing file (like `test_inference_similarity.py`) or create a new one if testing a distinct category of models.
 
 2. **Define Model Parameters**: Define the configuration for the model you want to test. This includes model path, dimensions, inference steps, and other generation parameters. **Note:** Consider using lower `num_inference_steps` or reduced resolution (e.g., 480p instead of 720p) to keep test execution time reasonable, provided it doesn't compromise the test's ability to detect regression.
 
@@ -86,15 +81,10 @@ To add a new SSIM test, follow these steps:
    ```
 
 4. **Reference Videos**:
-   * When running the test for the first time (or when updating the reference), the test will fail because the reference video is missing. The generated video will be saved in `fastvideo/tests/ssim/generated_videos/<quality-tier>/<GPU>_reference_videos`.
+   * When running the test for the first time (or when updating the reference), the test will fail because the reference video is missing. The generated video will be saved in `fastvideo/tests/ssim/generated_videos`.
    * Inspect the generated video to ensure it meets quality expectations.
-   * Move the generated video to the appropriate quality/GPU reference folder:
-     `fastvideo/tests/ssim/reference_videos/<quality-tier>/<GPU>_reference_videos/<Model>/<Backend>/`.
-   * You can use the helper CLI to copy generated videos into a reference folder:
-     `python fastvideo/tests/ssim/reference_videos_cli.py copy-local --quality-tier default --reference-dir fastvideo/tests/ssim/reference_videos/default/L40S_reference_videos`
-   * Upload/download can target both quality tiers and specific GPU folders:
-     `python fastvideo/tests/ssim/reference_videos_cli.py upload --quality-tier all`
-     `python fastvideo/tests/ssim/reference_videos_cli.py download --quality-tier full_quality --device-folder H200_reference_videos`
+   * Move the generated video to the appropriate reference folder: `fastvideo/tests/ssim/<GPU>_reference_videos/<Model>/<Backend>/`.
+   * You can use the helper script `update_reference_videos.sh` to automate copying videos from `generated_videos` to `L40S_reference_videos`. Note: Check the script to ensure paths match your environment (it defaults to `L40S_reference_videos`).
 
 ### Running Tests Locally
 
@@ -106,113 +96,36 @@ pytest fastvideo/tests/ssim/ -vs
 
 Ensure you have the necessary GPUs available as defined in your test parameters.
 
-## CI Integration
+## Modal Workflow
 
-FastVideo uses [Modal](https://modal.com/) for running tests in a CI environment. The
-workflow scripts are located in `fastvideo/tests/modal/`.
-
-### Buildkite Pipeline
-
-Tests are orchestrated by Buildkite (`.buildkite/pipeline.yml`) and executed on Modal GPU
-instances. The pipeline runs in two modes:
-
-**Fastcheck** — runs on every PR push, path-filtered. Only tests for the components you
-changed are triggered. Tests run in parallel.
-
-**Full Suite** — runs when a PR enters the Merge Queue (or when triggered manually via
-`/test full`). Covers SSIM regression, training, distillation, inference, and performance.
+FastVideo uses [Modal](https://modal.com/) for running tests in a CI environment. The workflow scripts are located in `fastvideo/tests/modal/`.
 
 ### `pr_test.py`
 
-The main entry point for CI tests is `fastvideo/tests/modal/pr_test.py`. This script defines
-Modal functions that execute the pytest suites on specific hardware (e.g., L40S, H100).
+The main entry point for CI tests is `fastvideo/tests/modal/pr_test.py`. This script defines Modal functions that execute the pytest suites on specific hardware (e.g., L40S, H100).
 
 ### Updating Modal Configuration
 
 If you add a new test that requires:
-
 * **Different GPU Hardware**: You may need to change the `@app.function(gpu=...)` decorator.
 * **Longer Execution Time**: Increase the `timeout` parameter.
-* **New Environment Variables/Secrets**: Add them to `secrets=[...]` or the image
-  environment. For example, if your model is gated on Hugging Face, ensure `HF_API_KEY`
-  is passed.
+* **New Environment Variables/Secrets**: Add them to `secrets=[...]` or the image environment. For example, if your model is gated on Hugging Face, ensure `HF_API_KEY` is passed.
 
-For SSIM tests, use `fastvideo/tests/modal/ssim_test.py`:
+For SSIM tests, the `run_ssim_tests` function in `pr_test.py` currently runs:
 
-```bash
-python -m modal run fastvideo/tests/modal/ssim_test.py::run_ssim_tests
+```python
+@app.function(gpu="L40S:2", image=image, timeout=2700, secrets=[modal.Secret.from_dict({"HF_API_KEY": os.environ.get("HF_API_KEY", "")})])
+def run_ssim_tests():
+    run_test("hf auth login --token $HF_API_KEY && pytest ./fastvideo/tests/ssim -vs")
 ```
 
-Target specific SSIM files/models:
-
-```bash
-python -m modal run fastvideo/tests/modal/ssim_test.py::run_ssim_tests \
-  --test-files test_wan_t2v_similarity.py \
-  --model-ids Wan2.1-T2V-1.3B-Diffusers
-```
-
-If HF token env vars are not set (`HF_API_KEY` / `HUGGINGFACE_HUB_TOKEN` /
-`HF_TOKEN`), the local entrypoint fails fast. To export raw `generated_videos`
-from Modal to the shared volume:
-
-```bash
-python -m modal run fastvideo/tests/modal/ssim_test.py::run_ssim_tests \
-  --sync-generated-to-volume
-```
-
-The raw export path is quality-tiered:
-
-* default params: `ssim_generated_videos/default/<subdir>/generated_videos`
-* full-quality params: `ssim_generated_videos/full_quality/<subdir>/generated_videos`
-
-The printed `modal volume get` command also downloads into a quality-specific
-local directory under `./generated_videos_modal/<quality-tier>`.
-
-To turn downloaded Modal outputs into local reference videos, use the matching
-quality tier with `copy-local`, for example:
-
-```bash
-python fastvideo/tests/ssim/reference_videos_cli.py copy-local \
-  --quality-tier full_quality \
-  --generated-dir ./generated_videos_modal/full_quality/L40S_reference_videos \
-  --device-folder L40S_reference_videos
-```
+If your new test file is inside `fastvideo/tests/ssim`, it will automatically be picked up by this command. However, ensure that the `gpu="L40S:2"` configuration is sufficient for your model. If your model requires more GPUs (e.g., 4 or 8), you might need to create a separate Modal function or update the existing one.
 
 ### Workflow Scripts
 
-The shell script that triggers tests in CI is `.buildkite/scripts/pr_test.sh`. If you add
-a new test category (e.g., a new folder outside of `ssim`), you will need to:
-
+The shell script that triggers these tests in the CI pipeline is located at `.buildkite/scripts/pr_test.sh`. If you add a new test category (e.g., a new folder outside of `ssim`), you will need to:
 1. Add a new function in `fastvideo/tests/modal/pr_test.py`.
 2. Add a new case in `.buildkite/scripts/pr_test.sh` to handle the new test type.
 
 !!! note
-    If you are a maintainer, update the workflow script in Buildkite after merging. Otherwise,
-    ask a maintainer for help.
-
-## Triggering Tests via Slash Commands
-
-Maintainers and contributors with write access can trigger individual test suites directly
-from a PR comment. The workflow reacts with a 🚀 emoji to confirm the command was received.
-
-```
-/test ssim              # SSIM regression tests
-/test training          # Training pipeline tests
-/test lora-training     # LoRA training tests
-/test lora-inference    # LoRA inference tests
-/test distillation      # DMD distillation tests
-/test self-forcing      # Self-Forcing tests
-/test vsa               # VSA training tests
-/test vmoba             # VMoBA inference tests
-/test performance       # Performance benchmarks
-/test api               # API server integration tests
-/test encoder           # Encoder component tests (Fastcheck)
-/test vae               # VAE component tests (Fastcheck)
-/test transformer       # Transformer / DiT tests (Fastcheck)
-/test kernel            # CUDA kernel tests (Fastcheck)
-/test unit              # Unit tests (Fastcheck)
-/test full              # Entire Full Suite
-/test fastcheck         # Entire Fastcheck suite
-```
-
-See [CI Architecture](ci_architecture.md) for the complete reference.
+    If you are a maintainer, you'll need to finally manually update the workflow script in Buildkite. Otherwise, a maintainer will help you update.

examples/train/README.md

@@ -1,93 +0,0 @@
-# Training
-
-## Quick Start
-
-### Single-node
-
-```bash
-bash examples/train/run.sh <config.yaml> [--dotted.key value ...]
-```
-
-The script auto-detects available GPUs. Override with `NUM_GPUS`:
-
-```bash
-NUM_GPUS=4 bash examples/train/run.sh \
-    examples/train/configs/fine_tuning/wan/t2v.yaml
-```
-
-### Multi-node (Slurm)
-
-```bash
-bash examples/train/run_slurm.sh <config.yaml> <num_nodes> [--dotted.key value ...]
-```
-
-```bash
-bash examples/train/run_slurm.sh \
-    examples/train/configs/fine_tuning/wan/t2v.yaml 4 \
-    --training.distributed.hsdp_shard_dim 8 \
-    --training.distributed.hsdp_replicate_dim 4
-```
-
-Slurm environment variables:
-
-| Variable | Default | Description |
-|---|---|---|
-| `PARTITION` | `main` | Slurm partition |
-| `NUM_GPUS` | `8` | GPUs per node |
-| `CPUS_PER_TASK` | `128` | CPUs per task |
-| `MEM` | `1440G` | Memory per node |
-| `EXCLUDE` | `""` | Nodes to exclude |
-
-### Overriding config values
-
-Any config field can be overridden from the command line using dotted keys:
-
-```bash
-bash examples/train/run.sh examples/train/configs/fine_tuning/wan/t2v.yaml \
-    --training.optimizer.learning_rate 1e-5 \
-    --training.loop.max_train_steps 1000 \
-    --training.checkpoint.output_dir outputs/my_experiment
-```
-
-### Resuming from a checkpoint
-
-```bash
-bash examples/train/run.sh examples/train/configs/fine_tuning/wan/t2v.yaml \
-    --training.checkpoint.resume_from_checkpoint outputs/my_experiment/checkpoint-500
-```
-
-## W&B Logging
-
-Training metrics and validation videos are logged to
-[Weights & Biases](https://wandb.ai). Set your API key before launching:
-
-```bash
-export WANDB_API_KEY=your_key_here
-```
-
-To disable W&B (e.g. for local debugging):
-
-```bash
-export WANDB_MODE=offline
-```
-
-The project name and run name are set in the config under `training.tracker`:
-
-```yaml
-training:
-  tracker:
-    project_name: my_project
-    run_name: my_run
-```
-
-## Directory Layout
-
-```
-examples/train/
-├── configs/      # Single-step training configs (by method and model)
-├── scenario/     # Multi-step end-to-end training pipelines
-├── run.sh        # Single-node launcher
-└── run_slurm.sh  # Multi-node Slurm launcher
-```
-
-See `configs/README.md` and `scenario/README.md` for details.

examples/train/configs/README.md

@@ -1,21 +0,0 @@
-# Training Configs
-
-Single-step training configurations organized by method and model.
-
-```
-configs/
-├── fine_tuning/          # Standard finetuning and DFSFT
-├── distribution_matching/  # DMD2 and Self-Forcing
-├── knowledge_distillation/ # KD from teacher to student
-└── example.yaml          # Annotated reference config with all fields
-```
-
-Each method directory contains per-model subdirectories (e.g. `wan/`, `hunyuan/`).
-
-Launch any config with:
-
-```bash
-bash examples/train/run.sh examples/train/configs/<method>/<model>/<config>.yaml
-```
-
-For multi-step training pipelines, see `examples/train/scenario/`.

examples/train/configs/fine_tuning/hunyuan/t2v.yaml

@@ -1,72 +0,0 @@
-# HunyuanVideo T2V finetune config.
-#
-# Data must be preprocessed with Hunyuan VAE + dual text encoders
-# (LLaMA + CLIP) into parquet format before training.
-# See scripts/preprocess/ for preprocessing pipeline.
-
-models:
-  student:
-    _target_: fastvideo.train.models.hunyuan.HunyuanModel
-    init_from: hunyuanvideo-community/HunyuanVideo
-    trainable: true
-
-method:
-  _target_: fastvideo.train.methods.fine_tuning.finetune.FineTuneMethod
-
-training:
-  distributed:
-    num_gpus: 8
-    sp_size: 1
-    tp_size: 1
-    hsdp_replicate_dim: 1
-    hsdp_shard_dim: 8
-
-  data:
-    data_path: data/hunyuan_training_data
-    dataloader_num_workers: 4
-    train_batch_size: 1
-    training_cfg_rate: 0.0
-    seed: 1000
-    num_latent_t: 32
-    num_height: 720
-    num_width: 1280
-    num_frames: 125
-
-  optimizer:
-    learning_rate: 1.0e-5
-    betas: [0.9, 0.999]
-    weight_decay: 0.01
-    lr_scheduler: constant
-    lr_warmup_steps: 0
-
-  loop:
-    max_train_steps: 5000
-    gradient_accumulation_steps: 1
-
-  checkpoint:
-    output_dir: outputs/hunyuan_finetune
-    training_state_checkpointing_steps: 500
-    checkpoints_total_limit: 3
-    resume_from_checkpoint: latest
-
-  tracker:
-    project_name: fastvideo_hunyuan
-    run_name: hunyuan_finetune
-
-  model:
-    enable_gradient_checkpointing_type: full
-
-callbacks:
-  grad_clip:
-    _target_: fastvideo.train.callbacks.grad_clip.GradNormClipCallback
-    max_grad_norm: 1.0
-  validation:
-    _target_: fastvideo.train.callbacks.validation.ValidationCallback
-    pipeline_target: fastvideo.pipelines.basic.hunyuan.hunyuan_pipeline.HunyuanVideoPipeline
-    dataset_file: data/hunyuan_training_data/validation_prompts.json
-    every_steps: 100
-    sampling_steps: [50]
-    guidance_scale: 1.0
-
-pipeline:
-  flow_shift: 7

examples/train/configs/fine_tuning/wan/t2v.yaml

@@ -1,68 +0,0 @@
-# Wan 2.1 T2V 1.3B finetune.
-
-models:
-  student:
-    _target_: fastvideo.train.models.wan.WanModel
-    init_from: Wan-AI/Wan2.1-T2V-1.3B-Diffusers
-    trainable: true
-
-method:
-  _target_: fastvideo.train.methods.fine_tuning.finetune.FineTuneMethod
-
-training:
-  distributed:
-    num_gpus: 8
-    sp_size: 1
-    tp_size: 1
-    hsdp_replicate_dim: 8
-    hsdp_shard_dim: 1
-
-  data:
-    data_path: data/preprocessed
-    dataloader_num_workers: 4
-    train_batch_size: 1
-    training_cfg_rate: 0.0
-    seed: 1000
-    num_latent_t: 20
-    num_height: 448
-    num_width: 832
-    num_frames: 77
-
-  optimizer:
-    learning_rate: 1.0e-6
-    betas: [0.9, 0.999]
-    weight_decay: 0.01
-    lr_scheduler: constant
-    lr_warmup_steps: 0
-
-  loop:
-    max_train_steps: 4000
-    gradient_accumulation_steps: 1
-
-  checkpoint:
-    output_dir: outputs/wan_finetune
-    training_state_checkpointing_steps: 100
-    checkpoints_total_limit: 3
-    resume_from_checkpoint: latest
-
-  tracker:
-    project_name: fastvideo
-    run_name: wan_finetune
-
-  model:
-    enable_gradient_checkpointing_type: full
-
-callbacks:
-  grad_clip:
-    _target_: fastvideo.train.callbacks.grad_clip.GradNormClipCallback
-    max_grad_norm: 1.0
-  validation:
-    _target_: fastvideo.train.callbacks.validation.ValidationCallback
-    pipeline_target: fastvideo.pipelines.basic.wan.wan_pipeline.WanPipeline
-    dataset_file: data/validation_prompts.json
-    every_steps: 100
-    sampling_steps: [50]
-    guidance_scale: 5.0
-
-pipeline:
-  flow_shift: 3

examples/train/configs/kdtest/decode_kd_cache.py

@@ -4,7 +4,7 @@ Uses fastvideo's AutoencoderKLWan and its built-in denormalization
 (latent * std + mean, matching normalize_dit_input's inverse).
 
 Usage:
-    python examples/train/scenario/ode_init_self_forcing_wan_causal/decode_kd_cache.py \
+    python examples/train/configs/kdtest/decode_kd_cache.py \
         --cache_dir data/kd_test_cache_small \
         --out_dir   data/kd_test_cache_small/video \
         --fps 16

examples/train/configs/kdtest/kd_causal.yaml

@@ -5,7 +5,7 @@
 #
 # Step 1 — Run KD training:
 #   torchrun --nproc_per_node=1 -m fastvideo.train.entrypoint.train \
-#       --config examples/train/scenario/ode_init_self_forcing_wan_causal/step1_kd.yaml
+#       --config examples/train/configs/kdtest/kd_causal.yaml
 #
 # Step 2 — Export ode_init checkpoint:
 #   python -m fastvideo.train.entrypoint.dcp_to_diffusers \
@@ -15,7 +15,7 @@
 #
 # Step 3 — Run Self-Forcing with the exported ode_init:
 #   torchrun --nproc_per_node=1 -m fastvideo.train.entrypoint.train \
-#       --config examples/train/scenario/ode_init_self_forcing_wan_causal/step3_self_forcing.yaml
+#       --config examples/train/configs/kdtest/self_forcing_causal.yaml
 #
 # Success criteria:
 #   - Cache generates 16 .pt files under data/kd_test_cache/samples/

examples/train/configs/kdtest/self_forcing_causal.yaml

@@ -1,7 +1,7 @@
 # Self-Forcing Test: Causal Wan 1.3B
 #
 # Smoke-test for SelfForcingMethod on a tiny 16-sample dataset.
-# Run this AFTER step1_kd.yaml + dcp_to_diffusers export.
+# Run this AFTER kd_causal.yaml + dcp_to_diffusers export.
 #
 # Prerequisite — export ode_init from the KD run:
 #   python -m fastvideo.train.entrypoint.dcp_to_diffusers \
@@ -11,7 +11,7 @@
 #
 # Then run this config:
 #   torchrun --nproc_per_node=1 -m fastvideo.train.entrypoint.train \
-#       --config examples/train/scenario/ode_init_self_forcing_wan_causal/step3_self_forcing.yaml
+#       --config examples/train/configs/kdtest/self_forcing_causal.yaml
 #
 # Success criteria:
 #   - generator_loss and critic_loss both finite at step 0
@@ -100,7 +100,7 @@ callbacks:
     max_grad_norm: 1.0
   validation:
     pipeline_target: fastvideo.pipelines.basic.wan.wan_causal_dmd_pipeline.WanCausalDMDPipeline
-    dataset_file: examples/train/scenario/ode_init_self_forcing_wan_causal/validation.json
+    dataset_file: examples/train/configs/kdtest/validation.json
     every_steps: 20
     sampling_steps: [4]
     sampling_timesteps: [1000, 750, 500, 250]

examples/train/configs/kdtest/usage.md

@@ -0,0 +1,8 @@
+# Step 1 — KD training                                                                                      
+  bash examples/train/run.sh examples/train/configs/kdtest/kd_causal.yaml                                     
+                                                                                                              
+  # Step 2 — Export ode_init                                                                                  
+python -m fastvideo.train.entrypoint.dcp_to_diffusers --role student --checkpoint outputs/kdtest/kd_causal/checkpoint-300 --output-dir outputs/kdtest/kd_ode_init
+                                                                                                              
+  # Step 3 — Self-Forcing                                                                                     
+  bash examples/train/run.sh examples/train/configs/kdtest/self_forcing_causal.yaml

examples/train/configs/overfit_hunyuan_t2v.yaml

@@ -1,78 +0,0 @@
-# HunyuanVideo T2V overfitting test config.
-#
-# Overfits on 3 short videos (480x832, 77 frames) to verify the
-# training pipeline works end-to-end.
-#
-# Preprocess data first:
-#   CUDA_VISIBLE_DEVICES=0 python scripts/preprocess_hunyuan_overfit.py
-#
-# Run:
-#   bash examples/train/run.sh examples/train/configs/overfit_hunyuan_t2v.yaml
-
-models:
-  student:
-    _target_: fastvideo.train.models.hunyuan.HunyuanModel
-    init_from: hunyuanvideo-community/HunyuanVideo
-    trainable: true
-    enable_gradient_checkpointing_type: full
-
-method:
-  _target_: fastvideo.train.methods.fine_tuning.finetune.FineTuneMethod
-
-training:
-  distributed:
-    num_gpus: 8
-    sp_size: 1
-    tp_size: 1
-    hsdp_replicate_dim: 1
-    hsdp_shard_dim: 8
-
-  data:
-    data_path: data/hunyuan_overfit_preprocessed
-    dataloader_num_workers: 0
-    train_batch_size: 1
-    training_cfg_rate: 0.0
-    seed: 42
-    num_latent_t: 20
-    num_height: 480
-    num_width: 832
-    num_frames: 77
-
-  optimizer:
-    learning_rate: 5.0e-5
-    betas: [0.9, 0.999]
-    weight_decay: 0.0
-    lr_scheduler: constant
-    lr_warmup_steps: 0
-
-  loop:
-    max_train_steps: 300
-    gradient_accumulation_steps: 1
-
-  checkpoint:
-    output_dir: outputs/hunyuan_overfit
-    training_state_checkpointing_steps: 50
-    checkpoints_total_limit: 2
-
-  tracker:
-    project_name: fastvideo_hunyuan
-    run_name: hunyuan_overfit
-
-  model:
-    precondition_outputs: true
-    enable_gradient_checkpointing_type: full
-
-callbacks:
-  grad_clip:
-    _target_: fastvideo.train.callbacks.grad_clip.GradNormClipCallback
-    max_grad_norm: 1.0
-  validation:
-    _target_: fastvideo.train.callbacks.validation.ValidationCallback
-    pipeline_target: fastvideo.pipelines.basic.hunyuan.hunyuan_pipeline.HunyuanVideoPipeline
-    dataset_file: data/hunyuan_overfit_preprocessed/validation_prompts.json
-    every_steps: 50
-    sampling_steps: [50]
-    guidance_scale: 1.0
-
-pipeline:
-  flow_shift: 7

examples/train/run.sh

@@ -5,12 +5,12 @@
 #   bash examples/train/run.sh <config.yaml> [--dotted.key value ...]
 #
 # Examples:
-#   bash examples/train/run.sh examples/train/configs/fine_tuning/wan/t2v.yaml
-#   bash examples/train/run.sh examples/train/configs/distribution_matching/wan/dmd2_t2v.yaml --dry-run
-#   bash examples/train/run.sh examples/train/configs/distribution_matching/wan/dmd2_t2v.yaml \
+#   bash examples/train/run.sh examples/train/finetune_wan2.1_t2v_1.3B_vsa_phase3.4_0.9sparsity.yaml
+#   bash examples/train/run.sh examples/train/distill_wan2.1_t2v_1.3B_dmd2.yaml --dry-run
+#   bash examples/train/run.sh examples/train/distill_wan2.1_t2v_1.3B_dmd2.yaml \
 #       --training.distributed.num_gpus 4 \
 #       --training.optimizer.learning_rate 1e-5
-#   bash examples/train/run.sh examples/train/configs/distribution_matching/wan/dmd2_t2v.yaml \
+#   bash examples/train/run.sh examples/train/distill_wan2.1_t2v_1.3B_dmd2.yaml \
 #       --training.checkpoint.resume_from_checkpoint outputs/my_run/checkpoint-1000
 #
 # Logs are written to logs/<config_name>_<timestamp>.log (and also printed to stdout).

examples/train/run_slurm.sh

@@ -6,7 +6,7 @@
 #
 # Examples:
 #   bash examples/train/run_slurm.sh examples/train/configs/example.yaml 2
-#   bash examples/train/run_slurm.sh examples/train/configs/distribution_matching/wan/dmd2_t2v.yaml 4 \
+#   bash examples/train/run_slurm.sh examples/train/configs/distill_wan2.1_t2v_1.3B_dmd2.yaml 4 \
 #       --training.optimizer.learning_rate 1e-5
 #   bash examples/train/run_slurm.sh examples/train/configs/example.yaml 8 \
 #       --training.checkpoint.resume_from_checkpoint outputs/my_run/checkpoint-1000

examples/train/scenario/README.md

@@ -1,13 +0,0 @@
-# Training Scenarios
-
-End-to-end multi-step training pipelines. Each scenario directory contains
-all configs, scripts, and data needed to run a complete workflow.
-
-```
-scenario/
-└── ode_init_self_forcing_wan_causal/   # KD → export → Self-Forcing
-```
-
-See the `usage.md` inside each scenario for step-by-step instructions.
-
-For single-step configs, see `examples/train/configs/`.

examples/train/scenario/ode_init_self_forcing_wan_causal/step2_export.sh

@@ -1,21 +0,0 @@
-#!/usr/bin/env bash
-# Step 2 — Export ODE-init weights from KD checkpoint to diffusers format.
-#
-# Usage:
-#   bash examples/train/scenario/ode_init_self_forcing_wan_causal/step2_export.sh
-
-set -euo pipefail
-
-CHECKPOINT_DIR="${1:-outputs/kdtest/kd_causal/checkpoint-300}"
-OUTPUT_DIR="${2:-outputs/kdtest/kd_ode_init}"
-
-echo "Exporting ODE-init weights..."
-echo "  checkpoint: ${CHECKPOINT_DIR}"
-echo "  output:     ${OUTPUT_DIR}"
-
-python -m fastvideo.train.entrypoint.dcp_to_diffusers \
-    --role student \
-    --checkpoint "${CHECKPOINT_DIR}" \
-    --output-dir "${OUTPUT_DIR}"
-
-echo "Done. Use the exported weights in step3_self_forcing.yaml."

examples/train/scenario/ode_init_self_forcing_wan_causal/usage.md

@@ -1,19 +0,0 @@
-# ODE-Init Self-Forcing: Wan 2.1 Causal 1.3B
-
-End-to-end scenario that trains a causal Wan 1.3B model via
-ODE-init knowledge distillation followed by self-forcing.
-
-## Steps
-
-```bash
-# Step 1 — KD training (produces ODE-init checkpoint)
-bash examples/train/run.sh \
-    examples/train/scenario/ode_init_self_forcing_wan_causal/step1_kd.yaml
-
-# Step 2 — Export ODE-init weights to diffusers format
-bash examples/train/scenario/ode_init_self_forcing_wan_causal/step2_export.sh
-
-# Step 3 — Self-forcing with the exported ODE-init
-bash examples/train/run.sh \
-    examples/train/scenario/ode_init_self_forcing_wan_causal/step3_self_forcing.yaml
-```

fastvideo/configs/models/dits/__init__.py

@@ -7,9 +7,8 @@ from fastvideo.configs.models.dits.longcat import LongCatVideoConfig
 from fastvideo.configs.models.dits.ltx2 import LTX2VideoConfig
 from fastvideo.configs.models.dits.wanvideo import WanVideoConfig
 from fastvideo.configs.models.dits.hyworld import HYWorldConfig
-from fastvideo.configs.models.dits.kandinsky5 import Kandinsky5VideoConfig
 
 __all__ = [
     "HunyuanVideoConfig", "HunyuanVideo15Config", "HunyuanGameCraftConfig", "WanVideoConfig", "CosmosVideoConfig",
-    "Cosmos25VideoConfig", "LongCatVideoConfig", "LTX2VideoConfig", "HYWorldConfig", "Kandinsky5VideoConfig"
+    "Cosmos25VideoConfig", "LongCatVideoConfig", "LTX2VideoConfig", "HYWorldConfig"
 ]

fastvideo/configs/models/dits/kandinsky5.py

@@ -1,65 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-from dataclasses import dataclass, field
-
-from fastvideo.configs.models.dits.base import DiTArchConfig, DiTConfig
-
-
-@dataclass
-class Kandinsky5ArchConfig(DiTArchConfig):
-    _fsdp_shard_conditions: list = field(default_factory=lambda: [
-        lambda n, m:
-        ("text_transformer_blocks" in n or "visual_transformer_blocks" in n) and n.split(".")[-1].isdigit()
-    ])
-
-    # Native FastVideo implementation uses the same parameter names as diffusers
-    # except FFN internals: Diffusers FFN uses `in_layer/out_layer`, while
-    # FastVideo uses MLP `fc_in/fc_out`.
-    param_names_mapping: dict = field(
-        default_factory=lambda: {
-            r"^(.*feed_forward)\.in_layer\.(weight|bias)$": r"\1.mlp.fc_in.\2",
-            r"^(.*feed_forward)\.out_layer\.(weight|bias)$": r"\1.mlp.fc_out.\2",
-        })
-
-    reverse_param_names_mapping: dict = field(default_factory=lambda: {})
-
-    # Diffusers Kandinsky5Transformer3DModel config fields.
-    in_visual_dim: int = 4
-    in_text_dim: int = 3584
-    in_text_dim2: int = 768
-    time_dim: int = 512
-    out_visual_dim: int = 4
-    patch_size: tuple[int, int, int] = (1, 2, 2)
-    model_dim: int = 2048
-    ff_dim: int = 5120
-    num_text_blocks: int = 2
-    num_visual_blocks: int = 32
-    axes_dims: tuple[int, int, int] = (16, 24, 24)
-    visual_cond: bool = False
-    attention_type: str = "regular"
-    attention_causal: bool | None = None
-    attention_local: bool | None = None
-    attention_glob: bool | None = None
-    attention_window: int | None = None
-    attention_P: float | None = None
-    attention_wT: int | None = None
-    attention_wW: int | None = None
-    attention_wH: int | None = None
-    attention_add_sta: bool | None = None
-    attention_method: str | None = None
-
-    def __post_init__(self):
-        super().__post_init__()
-        head_dim = sum(self.axes_dims)
-        if self.model_dim % head_dim != 0:
-            raise ValueError(f"model_dim ({self.model_dim}) must be divisible by head_dim ({head_dim})")
-        self.hidden_size = self.model_dim
-        self.num_attention_heads = self.model_dim // head_dim
-        self.in_channels = self.in_visual_dim
-        self.out_channels = self.out_visual_dim
-        self.num_channels_latents = self.in_visual_dim
-
-
-@dataclass
-class Kandinsky5VideoConfig(DiTConfig):
-    arch_config: DiTArchConfig = field(default_factory=Kandinsky5ArchConfig)
-    prefix: str = "Kandinsky5"

fastvideo/dataset/dataloader/record_schema.py

@@ -72,12 +72,9 @@ def i2v_record_creator(batch: PreprocessBatch) -> list[dict[str, Any]]:
     """Create a record for the Parquet dataset with CLIP features."""
     records = basic_t2v_record_creator(batch)
 
-    if len(batch.image_embeds) > 0:
-        assert len(batch.image_embeds
-                   ) == 1, "image embedding should be a single tensor"
-        image_embeds = batch.image_embeds[0]
-    else:
-        image_embeds = None
+    assert len(
+        batch.image_embeds) == 1, "image embedding should be a single tensor"
+    image_embeds = batch.image_embeds[0]
     image_latent = batch.image_latent
     pil_image = batch.pil_image
 

fastvideo/dataset/validation_dataset.py

@@ -4,7 +4,6 @@ import os
 import pathlib
 
 import datasets
-import numpy as np
 from torch.utils.data import IterableDataset
 
 from fastvideo.distributed import (get_sp_world_size, get_world_rank,
@@ -161,36 +160,5 @@ class ValidationDataset(IterableDataset):
                 else:
                     sample["control_video"] = load_video(control_video_path)
 
-            if sample.get("action_path", None) is not None:
-                action_path = sample["action_path"]
-                action_path = os.path.join(self.dir, action_path)
-                sample["action_path"] = action_path
-                if not pathlib.Path(action_path).is_file():
-                    logger.warning("Action file %s does not exist.",
-                                   action_path)
-                else:
-                    action_data = np.load(action_path, allow_pickle=True)
-                    num_frames = sample["num_frames"]
-                    if isinstance(action_data,
-                                  np.ndarray) and action_data.dtype == object:
-                        action_data = action_data.item()
-                    if isinstance(action_data, dict):
-                        kb_length = len(action_data["keyboard"])
-                        ms_length = len(action_data["mouse"])
-                        assert kb_length >= num_frames, (
-                            f"keyboard length {kb_length} is smaller than "
-                            f"num_frames {num_frames} for {action_path}.")
-                        assert ms_length >= num_frames, (
-                            f"mouse length {ms_length} is smaller than "
-                            f"num_frames {num_frames} for {action_path}.")
-                        sample["keyboard_cond"] = action_data["keyboard"][:num_frames]
-                        sample["mouse_cond"] = action_data["mouse"][:num_frames]
-                    else:
-                        kb_length = len(action_data)
-                        assert kb_length >= num_frames, (
-                            f"keyboard length {kb_length} is smaller than "
-                            f"num_frames {num_frames} for {action_path}.")
-                        sample["keyboard_cond"] = action_data[:num_frames]
-
             sample = {k: v for k, v in sample.items() if v is not None}
             yield sample

fastvideo/models/dits/kandinsky5.py

@@ -1,761 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-from __future__ import annotations
-
-import math
-from dataclasses import dataclass
-from typing import Any
-
-import torch
-import torch.nn as nn
-import torch.nn.functional as F
-
-from fastvideo.attention import LocalAttention
-from fastvideo.configs.models.dits import Kandinsky5VideoConfig
-from fastvideo.layers.layernorm import LayerNormScaleShift
-from fastvideo.layers.linear import ReplicatedLinear
-from fastvideo.layers.mlp import MLP
-from fastvideo.models.dits.base import BaseDiT
-from fastvideo.platforms import AttentionBackendEnum
-
-FRACTAL_PIXEL_SIZE = 8
-_ARCH_CONFIG_DEFAULTS = Kandinsky5VideoConfig().arch_config
-
-
-def _build_rotary_freqs(dim: int, max_period: float) -> torch.Tensor:
-    return torch.exp(-math.log(max_period) * torch.arange(
-        start=0, end=dim, dtype=torch.float32) / dim)
-
-
-def local_patching(x: torch.Tensor, shape: tuple[int, int, int, int],
-                   group_size: tuple[int, int, int],
-                   dim: int = 0) -> torch.Tensor:
-    batch_size, duration, height, width = shape
-    g1, g2, g3 = group_size
-    x = x.reshape(
-        *x.shape[:dim],
-        duration // g1,
-        g1,
-        height // g2,
-        g2,
-        width // g3,
-        g3,
-        *x.shape[dim + 3:],
-    )
-    x = x.permute(
-        *range(len(x.shape[:dim])),
-        dim,
-        dim + 2,
-        dim + 4,
-        dim + 1,
-        dim + 3,
-        dim + 5,
-        *range(dim + 6, len(x.shape)),
-    )
-    x = x.flatten(dim, dim + 2).flatten(dim + 1, dim + 3)
-    return x
-
-
-def local_merge(x: torch.Tensor, shape: tuple[int, int, int, int],
-                group_size: tuple[int, int, int],
-                dim: int = 0) -> torch.Tensor:
-    batch_size, duration, height, width = shape
-    g1, g2, g3 = group_size
-    x = x.reshape(
-        *x.shape[:dim],
-        duration // g1,
-        height // g2,
-        width // g3,
-        g1,
-        g2,
-        g3,
-        *x.shape[dim + 2:],
-    )
-    x = x.permute(
-        *range(len(x.shape[:dim])),
-        dim,
-        dim + 3,
-        dim + 1,
-        dim + 4,
-        dim + 2,
-        dim + 5,
-        *range(dim + 6, len(x.shape)),
-    )
-    x = x.flatten(dim, dim + 1).flatten(dim + 1, dim + 2).flatten(
-        dim + 2, dim + 3)
-    return x
-
-
-def fractal_flatten(x: torch.Tensor,
-                    rope: torch.Tensor,
-                    shape: tuple[int, int, int, int],
-                    block_mask: bool = False):
-    if block_mask:
-        pixel_size = FRACTAL_PIXEL_SIZE
-        x = local_patching(x, shape, (1, pixel_size, pixel_size), dim=1)
-        rope = local_patching(rope, shape, (1, pixel_size, pixel_size), dim=1)
-        x = x.flatten(1, 2)
-        rope = rope.flatten(1, 2)
-    else:
-        x = x.flatten(1, 3)
-        rope = rope.flatten(1, 3)
-    return x, rope
-
-
-def fractal_unflatten(x: torch.Tensor,
-                      shape: tuple[int, int, int, int],
-                      block_mask: bool = False):
-    if block_mask:
-        pixel_size = FRACTAL_PIXEL_SIZE
-        x = x.reshape(x.shape[0], -1, pixel_size**2, *x.shape[2:])
-        x = local_merge(x, shape, (1, pixel_size, pixel_size), dim=1)
-    else:
-        x = x.reshape(*shape, *x.shape[2:])
-    return x
-
-
-class Kandinsky5TimeEmbeddings(nn.Module):
-
-    def __init__(self, model_dim: int, time_dim: int,
-                 max_period: float = 10000.0):
-        super().__init__()
-        assert model_dim % 2 == 0
-        self.model_dim = model_dim
-        self.max_period = max_period
-        self.freqs = _build_rotary_freqs(self.model_dim // 2, self.max_period)
-        self.in_layer = ReplicatedLinear(model_dim, time_dim, bias=True)
-        self.activation = nn.SiLU()
-        self.out_layer = ReplicatedLinear(time_dim, time_dim, bias=True)
-
-    @torch.autocast(device_type="cuda", dtype=torch.float32)
-    def forward(self, time: torch.Tensor) -> torch.Tensor:
-        args = torch.outer(time, self.freqs.to(device=time.device))
-        time_embed = torch.cat([torch.cos(args), torch.sin(args)], dim=-1)
-        time_embed, _ = self.in_layer(time_embed)
-        time_embed = self.activation(time_embed)
-        time_embed, _ = self.out_layer(time_embed)
-        return time_embed
-
-
-class Kandinsky5TextEmbeddings(nn.Module):
-
-    def __init__(self, text_dim: int, model_dim: int):
-        super().__init__()
-        self.in_layer = ReplicatedLinear(text_dim, model_dim, bias=True)
-        self.norm = nn.LayerNorm(model_dim, elementwise_affine=True)
-
-    def forward(self, text_embed: torch.Tensor) -> torch.Tensor:
-        text_embed, _ = self.in_layer(text_embed)
-        return self.norm(text_embed).type_as(text_embed)
-
-
-class Kandinsky5VisualEmbeddings(nn.Module):
-
-    def __init__(self, visual_dim: int, model_dim: int,
-                 patch_size: tuple[int, int, int]):
-        super().__init__()
-        self.patch_size = patch_size
-        self.in_layer = ReplicatedLinear(math.prod(patch_size) * visual_dim,
-                                         model_dim)
-
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        batch_size, duration, height, width, dim = x.shape
-        x = (x.view(
-            batch_size,
-            duration // self.patch_size[0],
-            self.patch_size[0],
-            height // self.patch_size[1],
-            self.patch_size[1],
-            width // self.patch_size[2],
-            self.patch_size[2],
-            dim,
-        ).permute(0, 1, 3, 5, 2, 4, 6, 7).flatten(4, 7))
-        x, _ = self.in_layer(x)
-        return x
-
-
-class Kandinsky5RoPE1D(nn.Module):
-
-    def __init__(self,
-                 dim: int,
-                 max_pos: int = 1024,
-                 max_period: float = 10000.0):
-        super().__init__()
-        self.max_period = max_period
-        self.dim = dim
-        self.max_pos = max_pos
-        freq = _build_rotary_freqs(dim // 2, max_period)
-        pos = torch.arange(max_pos, dtype=freq.dtype)
-        self.register_buffer("args", torch.outer(pos, freq), persistent=False)
-
-    def forward(self, pos: torch.Tensor) -> torch.Tensor:
-        args = self.args[pos]
-        cosine = torch.cos(args)
-        sine = torch.sin(args)
-        rope = torch.stack([cosine, -sine, sine, cosine], dim=-1)
-        rope = rope.view(*rope.shape[:-1], 2, 2)
-        return rope.unsqueeze(-4)
-
-
-class Kandinsky5RoPE3D(nn.Module):
-
-    def __init__(self,
-                 axes_dims: tuple[int, int, int],
-                 max_pos: tuple[int, int, int] = (128, 128, 128),
-                 max_period: float = 10000.0):
-        super().__init__()
-        self.axes_dims = axes_dims
-        self.max_pos = max_pos
-        self.max_period = max_period
-
-        for i, (axes_dim, ax_max_pos) in enumerate(
-                zip(axes_dims, max_pos, strict=True)):
-            freq = _build_rotary_freqs(axes_dim // 2, max_period)
-            pos = torch.arange(ax_max_pos, dtype=freq.dtype)
-            self.register_buffer(f"args_{i}",
-                                 torch.outer(pos, freq),
-                                 persistent=False)
-
-    def forward(self,
-                shape: tuple[int, int, int, int],
-                pos: tuple[torch.Tensor, torch.Tensor, torch.Tensor]
-                | list[torch.Tensor],
-                scale_factor: tuple[float, float,
-                                    float] = (1.0, 1.0, 1.0)):
-        batch_size, duration, height, width = shape
-        args_t = self.args_0[pos[0]] / scale_factor[0]
-        args_h = self.args_1[pos[1]] / scale_factor[1]
-        args_w = self.args_2[pos[2]] / scale_factor[2]
-
-        args = torch.cat(
-            [
-                args_t.view(1, duration, 1, 1, -1).repeat(
-                    batch_size, 1, height, width, 1),
-                args_h.view(1, 1, height, 1, -1).repeat(
-                    batch_size, duration, 1, width, 1),
-                args_w.view(1, 1, 1, width, -1).repeat(
-                    batch_size, duration, height, 1, 1),
-            ],
-            dim=-1,
-        )
-        cosine = torch.cos(args)
-        sine = torch.sin(args)
-        rope = torch.stack([cosine, -sine, sine, cosine], dim=-1)
-        rope = rope.view(*rope.shape[:-1], 2, 2)
-        return rope.unsqueeze(-4)
-
-
-class Kandinsky5Modulation(nn.Module):
-
-    def __init__(self, time_dim: int, model_dim: int, num_params: int):
-        super().__init__()
-        self.activation = nn.SiLU()
-        self.out_layer = ReplicatedLinear(time_dim,
-                                          num_params * model_dim,
-                                          bias=True)
-        self.out_layer.weight.data.zero_()
-        self.out_layer.bias.data.zero_()
-
-    @torch.autocast(device_type="cuda", dtype=torch.float32)
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        x = self.activation(x)
-        x, _ = self.out_layer(x)
-        return x
-
-
-def _apply_rotary(x: torch.Tensor, rope: torch.Tensor) -> torch.Tensor:
-    orig_dtype = x.dtype
-    x_ = x.reshape(*x.shape[:-1], -1, 1, 2).to(torch.float32)
-    x_out = (rope * x_).sum(dim=-1)
-    return x_out.reshape(*x.shape).to(orig_dtype)
-
-
-class Kandinsky5Attention(nn.Module):
-
-    def __init__(
-        self,
-        num_channels: int,
-        head_dim: int,
-        supported_attention_backends: tuple[AttentionBackendEnum, ...] | None,
-        prefix: str = "",
-    ):
-        super().__init__()
-        assert num_channels % head_dim == 0
-        self.num_heads = num_channels // head_dim
-
-        self.to_query = ReplicatedLinear(num_channels,
-                                         num_channels,
-                                         bias=True,
-                                         prefix=f"{prefix}.to_query")
-        self.to_key = ReplicatedLinear(num_channels,
-                                       num_channels,
-                                       bias=True,
-                                       prefix=f"{prefix}.to_key")
-        self.to_value = ReplicatedLinear(num_channels,
-                                         num_channels,
-                                         bias=True,
-                                         prefix=f"{prefix}.to_value")
-        self.query_norm = nn.RMSNorm(head_dim)
-        self.key_norm = nn.RMSNorm(head_dim)
-        self.out_layer = ReplicatedLinear(num_channels,
-                                          num_channels,
-                                          bias=True,
-                                          prefix=f"{prefix}.out_layer")
-        self.local_attention = LocalAttention(
-            num_heads=self.num_heads,
-            head_size=head_dim,
-            causal=False,
-            supported_attention_backends=supported_attention_backends,
-        )
-
-    def forward(
-        self,
-        hidden_states: torch.Tensor,
-        encoder_hidden_states: torch.Tensor | None = None,
-        sparse_params: dict[str, Any] | None = None,
-        rotary_emb: torch.Tensor | None = None,
-    ) -> torch.Tensor:
-        query, _ = self.to_query(hidden_states)
-
-        if encoder_hidden_states is not None:
-            key, _ = self.to_key(encoder_hidden_states)
-            value, _ = self.to_value(encoder_hidden_states)
-            shape, cond_shape = query.shape[:-1], key.shape[:-1]
-            query = query.reshape(*shape, self.num_heads, -1)
-            key = key.reshape(*cond_shape, self.num_heads, -1)
-            value = value.reshape(*cond_shape, self.num_heads, -1)
-        else:
-            key, _ = self.to_key(hidden_states)
-            value, _ = self.to_value(hidden_states)
-            shape = query.shape[:-1]
-            query = query.reshape(*shape, self.num_heads, -1)
-            key = key.reshape(*shape, self.num_heads, -1)
-            value = value.reshape(*shape, self.num_heads, -1)
-
-        query = self.query_norm(query.float()).type_as(query)
-        key = self.key_norm(key.float()).type_as(key)
-
-        if rotary_emb is not None:
-            query = _apply_rotary(query, rotary_emb).type_as(query)
-            key = _apply_rotary(key, rotary_emb).type_as(key)
-
-        if sparse_params is not None:
-            raise NotImplementedError(
-                "Sparse attention is not yet supported for Kandinsky5 in FastVideo."
-            )
-
-        try:
-            hidden_states = self.local_attention(query, key, value)
-        except AssertionError as exc:
-            # LocalAttention requires pipeline forward context. Standalone
-            # parity tests call the model directly, so fallback to Torch SDPA.
-            if "Forward context is not set" not in str(exc):
-                raise
-            query = query.transpose(1, 2)
-            key = key.transpose(1, 2)
-            value = value.transpose(1, 2)
-            hidden_states = F.scaled_dot_product_attention(query,
-                                                           key,
-                                                           value,
-                                                           attn_mask=None,
-                                                           is_causal=False)
-            hidden_states = hidden_states.transpose(1, 2)
-        hidden_states = hidden_states.flatten(2)
-
-        hidden_states, _ = self.out_layer(hidden_states)
-        return hidden_states
-
-
-class Kandinsky5FeedForward(nn.Module):
-
-    def __init__(self, dim: int, ff_dim: int):
-        super().__init__()
-        self.mlp = MLP(dim, ff_dim, bias=False, act_type="gelu")
-
-    def forward(self, x: torch.Tensor) -> torch.Tensor:
-        return self.mlp(x)
-
-
-class Kandinsky5OutLayer(nn.Module):
-
-    def __init__(self, model_dim: int, time_dim: int, visual_dim: int,
-                 patch_size: tuple[int, int, int]):
-        super().__init__()
-        self.patch_size = patch_size
-        self.modulation = Kandinsky5Modulation(time_dim, model_dim, 2)
-        self.norm = nn.LayerNorm(model_dim,
-                                 eps=1e-5,
-                                 elementwise_affine=False)
-        self.out_layer = ReplicatedLinear(model_dim,
-                                          math.prod(patch_size) * visual_dim,
-                                          bias=True)
-
-    def forward(self, visual_embed: torch.Tensor,
-                time_embed: torch.Tensor) -> torch.Tensor:
-        shift, scale = torch.chunk(
-            self.modulation(time_embed).unsqueeze(dim=1), 2, dim=-1)
-        visual_embed = (
-            self.norm(visual_embed.float()) *
-            (scale.float()[:, None, None] + 1.0) +
-            shift.float()[:, None, None]).type_as(visual_embed)
-
-        x, _ = self.out_layer(visual_embed)
-
-        batch_size, duration, height, width, _ = x.shape
-        x = (x.view(
-            batch_size,
-            duration,
-            height,
-            width,
-            -1,
-            self.patch_size[0],
-            self.patch_size[1],
-            self.patch_size[2],
-        ).permute(0, 1, 5, 2, 6, 3, 7, 4).flatten(1, 2).flatten(
-            2, 3).flatten(3, 4))
-        return x
-
-
-class Kandinsky5TransformerEncoderBlock(nn.Module):
-
-    def __init__(self, model_dim: int, time_dim: int, ff_dim: int,
-                 head_dim: int,
-                 supported_attention_backends: tuple[AttentionBackendEnum, ...]
-                 | None = None,
-                 prefix: str = ""):
-        super().__init__()
-        self.text_modulation = Kandinsky5Modulation(time_dim, model_dim, 6)
-
-        self.self_attention_norm = LayerNormScaleShift(
-            model_dim,
-            norm_type="layer",
-            eps=1e-5,
-            elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
-        self.self_attention = Kandinsky5Attention(
-            model_dim,
-            head_dim,
-            supported_attention_backends=supported_attention_backends,
-            prefix=f"{prefix}.self_attention")
-
-        self.feed_forward_norm = LayerNormScaleShift(
-            model_dim,
-            norm_type="layer",
-            eps=1e-5,
-            elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
-        self.feed_forward = Kandinsky5FeedForward(model_dim, ff_dim)
-
-    def forward(self, x: torch.Tensor, time_embed: torch.Tensor,
-                rope: torch.Tensor) -> torch.Tensor:
-        self_attn_params, ff_params = torch.chunk(
-            self.text_modulation(time_embed).unsqueeze(dim=1), 2, dim=-1)
-        shift, scale, gate = torch.chunk(self_attn_params, 3, dim=-1)
-        out = self.self_attention_norm(x.float(),
-                                       shift=shift,
-                                       scale=scale,
-                                       convert_modulation_dtype=True).type_as(x)
-        out = self.self_attention(out, rotary_emb=rope)
-        x = (x.float() + gate.float() * out.float()).type_as(x)
-
-        ff_shift, ff_scale, ff_gate = torch.chunk(ff_params, 3, dim=-1)
-        out = self.feed_forward_norm(x.float(),
-                                     shift=ff_shift,
-                                     scale=ff_scale,
-                                     convert_modulation_dtype=True).type_as(x)
-        out = self.feed_forward(out)
-        x = (x.float() + ff_gate.float() * out.float()).type_as(x)
-
-        return x
-
-
-class Kandinsky5TransformerDecoderBlock(nn.Module):
-
-    def __init__(self, model_dim: int, time_dim: int, ff_dim: int,
-                 head_dim: int,
-                 supported_attention_backends: tuple[AttentionBackendEnum, ...]
-                 | None = None,
-                 prefix: str = ""):
-        super().__init__()
-        self.visual_modulation = Kandinsky5Modulation(time_dim, model_dim, 9)
-
-        self.self_attention_norm = LayerNormScaleShift(
-            model_dim,
-            norm_type="layer",
-            eps=1e-5,
-            elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
-        self.self_attention = Kandinsky5Attention(
-            model_dim,
-            head_dim,
-            supported_attention_backends=supported_attention_backends,
-            prefix=f"{prefix}.self_attention")
-
-        self.cross_attention_norm = LayerNormScaleShift(
-            model_dim,
-            norm_type="layer",
-            eps=1e-5,
-            elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
-        self.cross_attention = Kandinsky5Attention(
-            model_dim,
-            head_dim,
-            supported_attention_backends=supported_attention_backends,
-            prefix=f"{prefix}.cross_attention")
-
-        self.feed_forward_norm = LayerNormScaleShift(
-            model_dim,
-            norm_type="layer",
-            eps=1e-5,
-            elementwise_affine=False,
-            dtype=torch.float32,
-            compute_dtype=torch.float32)
-        self.feed_forward = Kandinsky5FeedForward(model_dim, ff_dim)
-
-    def forward(self, visual_embed: torch.Tensor, text_embed: torch.Tensor,
-                time_embed: torch.Tensor, rope: torch.Tensor,
-                sparse_params: dict[str, Any] | None) -> torch.Tensor:
-        self_attn_params, cross_attn_params, ff_params = torch.chunk(
-            self.visual_modulation(time_embed).unsqueeze(dim=1), 3, dim=-1)
-
-        self_shift, self_scale, self_gate = torch.chunk(self_attn_params,
-                                                        3,
-                                                        dim=-1)
-        visual_out = self.self_attention_norm(
-            visual_embed.float(),
-            shift=self_shift,
-            scale=self_scale,
-            convert_modulation_dtype=True,
-        ).type_as(visual_embed)
-        visual_out = self.self_attention(visual_out,
-                                         rotary_emb=rope,
-                                         sparse_params=sparse_params)
-        visual_embed = (visual_embed.float() +
-                        self_gate.float() * visual_out.float()).type_as(
-                            visual_embed)
-
-        cross_shift, cross_scale, cross_gate = torch.chunk(cross_attn_params,
-                                                           3,
-                                                           dim=-1)
-        visual_out = self.cross_attention_norm(
-            visual_embed.float(),
-            shift=cross_shift,
-            scale=cross_scale,
-            convert_modulation_dtype=True,
-        ).type_as(visual_embed)
-        visual_out = self.cross_attention(
-            visual_out, encoder_hidden_states=text_embed)
-        visual_embed = (visual_embed.float() +
-                        cross_gate.float() * visual_out.float()).type_as(
-                            visual_embed)
-
-        ff_shift, ff_scale, ff_gate = torch.chunk(ff_params, 3, dim=-1)
-        visual_out = self.feed_forward_norm(
-            visual_embed.float(),
-            shift=ff_shift,
-            scale=ff_scale,
-            convert_modulation_dtype=True,
-        ).type_as(visual_embed)
-        visual_out = self.feed_forward(visual_out)
-        visual_embed = (visual_embed.float() +
-                        ff_gate.float() * visual_out.float()).type_as(
-                            visual_embed)
-
-        return visual_embed
-
-
-@dataclass
-class Kandinsky5TransformerOutput:
-    sample: torch.Tensor
-
-
-class Kandinsky5Transformer3DModel(BaseDiT):
-    """
-    Native FastVideo implementation of Kandinsky5 Transformer.
-    """
-
-    _fsdp_shard_conditions = _ARCH_CONFIG_DEFAULTS._fsdp_shard_conditions
-    _compile_conditions = _ARCH_CONFIG_DEFAULTS._compile_conditions
-    param_names_mapping = _ARCH_CONFIG_DEFAULTS.param_names_mapping
-    reverse_param_names_mapping = _ARCH_CONFIG_DEFAULTS.reverse_param_names_mapping
-    lora_param_names_mapping = _ARCH_CONFIG_DEFAULTS.lora_param_names_mapping
-    _supported_attention_backends = _ARCH_CONFIG_DEFAULTS._supported_attention_backends
-
-    def __init__(self, config: Kandinsky5VideoConfig,
-                 hf_config: dict[str, Any]) -> None:
-        super().__init__(config=config, hf_config=hf_config)
-        arch = config.arch_config
-
-        head_dim = sum(arch.axes_dims)
-        self.in_visual_dim = arch.in_visual_dim
-        self.model_dim = arch.model_dim
-        self.patch_size = arch.patch_size
-        self.visual_cond = arch.visual_cond
-        self.attention_type = arch.attention_type
-
-        visual_embed_dim = (2 * arch.in_visual_dim +
-                            1) if arch.visual_cond else arch.in_visual_dim
-
-        self.time_embeddings = Kandinsky5TimeEmbeddings(
-            arch.model_dim, arch.time_dim)
-        self.text_embeddings = Kandinsky5TextEmbeddings(
-            arch.in_text_dim, arch.model_dim)
-        self.pooled_text_embeddings = Kandinsky5TextEmbeddings(
-            arch.in_text_dim2, arch.time_dim)
-        self.visual_embeddings = Kandinsky5VisualEmbeddings(
-            visual_embed_dim, arch.model_dim, arch.patch_size)
-
-        self.text_rope_embeddings = Kandinsky5RoPE1D(head_dim)
-        self.visual_rope_embeddings = Kandinsky5RoPE3D(arch.axes_dims)
-
-        self.text_transformer_blocks = nn.ModuleList([
-            Kandinsky5TransformerEncoderBlock(arch.model_dim, arch.time_dim,
-                                              arch.ff_dim,
-                                              head_dim,
-                                              self._supported_attention_backends,
-                                              prefix=f"{config.prefix}.text_transformer_blocks.{i}")
-            for i in range(arch.num_text_blocks)
-        ])
-        self.visual_transformer_blocks = nn.ModuleList([
-            Kandinsky5TransformerDecoderBlock(arch.model_dim, arch.time_dim,
-                                              arch.ff_dim,
-                                              head_dim,
-                                              self._supported_attention_backends,
-                                              prefix=f"{config.prefix}.visual_transformer_blocks.{i}")
-            for i in range(arch.num_visual_blocks)
-        ])
-
-        self.out_layer = Kandinsky5OutLayer(arch.model_dim, arch.time_dim,
-                                            arch.out_visual_dim,
-                                            arch.patch_size)
-        self.gradient_checkpointing = False
-
-        self.hidden_size = arch.hidden_size
-        self.num_attention_heads = arch.num_attention_heads
-        self.num_channels_latents = arch.num_channels_latents
-        self.__post_init__()
-
-    def forward(
-        self,
-        hidden_states: torch.Tensor,
-        encoder_hidden_states: torch.Tensor | list[torch.Tensor],
-        timestep: torch.Tensor,
-        encoder_hidden_states_image: torch.Tensor | list[torch.Tensor]
-        | None = None,
-        pooled_projections: torch.Tensor | None = None,
-        visual_rope_pos: tuple[torch.Tensor, torch.Tensor,
-                               torch.Tensor] | list[torch.Tensor] | None = None,
-        text_rope_pos: torch.Tensor | None = None,
-        scale_factor: tuple[float, float, float] = (1.0, 1.0, 1.0),
-        sparse_params: dict[str, Any] | None = None,
-        return_dict: bool = False,
-        **kwargs,
-    ) -> torch.Tensor | Kandinsky5TransformerOutput:
-        if isinstance(encoder_hidden_states, list):
-            encoder_hidden_states = encoder_hidden_states[0]
-
-        if pooled_projections is None:
-            if encoder_hidden_states_image is None:
-                raise ValueError(
-                    "pooled_projections must be provided for Kandinsky5.")
-            pooled_projections = encoder_hidden_states_image
-        if isinstance(pooled_projections, list):
-            pooled_projections = pooled_projections[0]
-
-        if visual_rope_pos is None or text_rope_pos is None:
-            raise ValueError(
-                "visual_rope_pos and text_rope_pos are required for Kandinsky5."
-            )
-
-        x = hidden_states
-        text_embed = self.text_embeddings(encoder_hidden_states)
-        time_embed = self.time_embeddings(timestep)
-        time_embed = time_embed + self.pooled_text_embeddings(pooled_projections)
-        visual_embed = self.visual_embeddings(x)
-        text_rope = self.text_rope_embeddings(text_rope_pos).unsqueeze(dim=0)
-
-        for text_transformer_block in self.text_transformer_blocks:
-            if torch.is_grad_enabled() and self.gradient_checkpointing:
-                text_embed = torch.utils.checkpoint.checkpoint(
-                    text_transformer_block,
-                    text_embed,
-                    time_embed,
-                    text_rope,
-                    use_reentrant=False)
-            else:
-                text_embed = text_transformer_block(text_embed, time_embed,
-                                                    text_rope)
-
-        visual_shape = visual_embed.shape[:-1]
-        visual_rope = self.visual_rope_embeddings(visual_shape, visual_rope_pos,
-                                                  scale_factor)
-        to_fractal = sparse_params[
-            "to_fractal"] if sparse_params is not None else False
-        visual_embed, visual_rope = fractal_flatten(visual_embed, visual_rope,
-                                                    visual_shape,
-                                                    block_mask=to_fractal)
-
-        for visual_transformer_block in self.visual_transformer_blocks:
-            if torch.is_grad_enabled() and self.gradient_checkpointing:
-                visual_embed = torch.utils.checkpoint.checkpoint(
-                    visual_transformer_block,
-                    visual_embed,
-                    text_embed,
-                    time_embed,
-                    visual_rope,
-                    sparse_params,
-                    use_reentrant=False)
-            else:
-                visual_embed = visual_transformer_block(
-                    visual_embed,
-                    text_embed,
-                    time_embed,
-                    visual_rope,
-                    sparse_params,
-                )
-
-        visual_embed = fractal_unflatten(visual_embed,
-                                         visual_shape,
-                                         block_mask=to_fractal)
-        x = self.out_layer(visual_embed, time_embed)
-
-        if return_dict:
-            return Kandinsky5TransformerOutput(sample=x)
-        return x
-
-    def materialize_non_persistent_buffers(self, device: torch.device,
-                                           dtype: torch.dtype | None = None
-                                           ) -> None:
-        if isinstance(self.time_embeddings.freqs,
-                      torch.Tensor) and self.time_embeddings.freqs.is_meta:
-            self.time_embeddings.freqs = _build_rotary_freqs(
-                self.time_embeddings.model_dim // 2,
-                self.time_embeddings.max_period).to(device=device)
-
-        if isinstance(self.text_rope_embeddings.args,
-                      torch.Tensor) and self.text_rope_embeddings.args.is_meta:
-            freq = _build_rotary_freqs(self.text_rope_embeddings.dim // 2,
-                                       self.text_rope_embeddings.max_period).to(
-                                           device=device)
-            pos = torch.arange(self.text_rope_embeddings.max_pos,
-                               dtype=freq.dtype,
-                               device=device)
-            self.text_rope_embeddings._buffers["args"] = torch.outer(pos, freq)
-
-        for i, (axes_dim,
-                ax_max_pos) in enumerate(
-                    zip(self.visual_rope_embeddings.axes_dims,
-                        self.visual_rope_embeddings.max_pos,
-                        strict=True)):
-            name = f"args_{i}"
-            buf = getattr(self.visual_rope_embeddings, name, None)
-            if isinstance(buf, torch.Tensor) and buf.is_meta:
-                freq = _build_rotary_freqs(
-                    axes_dim // 2, self.visual_rope_embeddings.max_period).to(
-                        device=device)
-                pos = torch.arange(ax_max_pos, dtype=freq.dtype, device=device)
-                self.visual_rope_embeddings._buffers[name] = torch.outer(
-                    pos, freq)

fastvideo/models/loader/fsdp_load.py

@@ -147,9 +147,6 @@ def maybe_load_fsdp_model(
         cpu_offload=cpu_offload,
         param_names_mapping=param_names_mapping_fn,
     )
-    if hasattr(model, "materialize_non_persistent_buffers"):
-        model.materialize_non_persistent_buffers(
-            device=device, dtype=default_dtype)
     for n, p in chain(model.named_parameters(), model.named_buffers()):
         if p.is_meta:
             raise RuntimeError(

fastvideo/models/registry.py

@@ -40,8 +40,6 @@ _TEXT_TO_VIDEO_DIT_MODELS = {
     "LTX2Transformer3DModel": ("dits", "ltx2", "LTX2Transformer3DModel"),
     "SD3Transformer2DModel": ("dits", "sd3", "SD3Transformer2DModel"),
     "LingBotWorldTransformer3DModel": ("dits", "lingbotworld", "LingBotWorldTransformer3DModel"),
-    "Kandinsky5Transformer3DModel":
-    ("dits", "kandinsky5", "Kandinsky5Transformer3DModel"),
 }
 
 _IMAGE_TO_VIDEO_DIT_MODELS = {

fastvideo/models/vaes/common.py

@@ -364,15 +364,10 @@ class ParallelTiledVAE(ABC):
             if i > 0:
                 tile = tile[:, :, 1:, :, :]
             row.append(tile)
-        blend_latent_num_frames = max(
-            1,
-            self.blend_num_frames // self.temporal_compression_ratio,
-        )
         result_row = []
         for i, tile in enumerate(row):
             if i > 0:
-                tile = self.blend_t(row[i - 1], tile,
-                                    blend_latent_num_frames)
+                tile = self.blend_t(row[i - 1], tile, self.blend_num_frames)
                 result_row.append(
                     tile[:, :, :tile_latent_stride_num_frames, :, :])
             else:

fastvideo/pipelines/preprocess/matrixgame/matrixgame_preprocess_pipeline.py

@@ -117,26 +117,56 @@ class PreprocessPipeline_MatrixGame(BasePreprocessPipeline):
         if "action_path" in valid_data and valid_data["action_path"]:
             keyboard_cond_list = []
             mouse_cond_list = []
+            num_bits = 6
             for action_path in valid_data["action_path"]:
                 if action_path:
                     action_data = np.load(action_path, allow_pickle=True)
                     if isinstance(action_data, np.ndarray) and action_data.dtype == np.dtype('O'):
                         action_dict = action_data.item()
                         if "keyboard" in action_dict:
-                            kb_length = len(action_dict["keyboard"])
-                            assert kb_length >= num_frames, (f"keyboard length {kb_length} is smaller than "
-                                                             f"num_frames {num_frames} for {action_path}.")
-                            keyboard_cond_list.append(action_dict["keyboard"][:num_frames].astype(np.float32))
+                            keyboard_raw = action_dict["keyboard"]
+                            # Convert 1D bit-flag values to 2D multi-hot encoding
+                            if isinstance(keyboard_raw, np.ndarray):
+                                if keyboard_raw.ndim == 1:
+                                    # [T] -> [T, num_bits]
+                                    T = len(keyboard_raw)
+                                    multi_hot = np.zeros((T, num_bits), dtype=np.float32)
+                                    action_values = keyboard_raw.astype(int)
+                                    for bit_idx in range(num_bits):
+                                        target_idx = (2 - (bit_idx % 3)) + 3 * (bit_idx // 3)
+                                        if target_idx < num_bits:
+                                            multi_hot[:, target_idx] = ((action_values >> bit_idx)
+                                                                        & 1).astype(np.float32)
+                                    keyboard_cond_list.append(multi_hot)
+                                else:
+                                    # If already 2D, pad to num_bits if necessary
+                                    k_data = keyboard_raw.astype(np.float32)
+                                    if k_data.ndim == 2 and k_data.shape[-1] < num_bits:
+                                        padding = np.zeros((k_data.shape[0], num_bits - k_data.shape[-1]),
+                                                           dtype=np.float32)
+                                        k_data = np.concatenate([k_data, padding], axis=-1)
+                                    keyboard_cond_list.append(k_data)
+                            else:
+                                keyboard_cond_list.append(keyboard_raw)
                         if "mouse" in action_dict:
-                            ms_length = len(action_dict["mouse"])
-                            assert ms_length >= num_frames, (f"mouse length {ms_length} is smaller than "
-                                                             f"num_frames {num_frames} for {action_path}.")
-                            mouse_cond_list.append(action_dict["mouse"][:num_frames])
+                            mouse_cond_list.append(action_dict["mouse"])
                     else:
-                        kb_length = len(action_data)
-                        assert kb_length >= num_frames, (f"keyboard length {kb_length} is smaller than "
-                                                         f"num_frames {num_frames} for {action_path}.")
-                        keyboard_cond_list.append(action_data.astype(np.float32)[:num_frames])
+                        if isinstance(action_data, np.ndarray) and action_data.ndim == 1:
+                            T = len(action_data)
+                            multi_hot = np.zeros((T, num_bits), dtype=np.float32)
+                            action_values = action_data.astype(int)
+                            for bit_idx in range(num_bits):
+                                target_idx = (2 - (bit_idx % 3)) + 3 * (bit_idx // 3)
+                                if target_idx < num_bits:
+                                    multi_hot[:, target_idx] = ((action_values >> bit_idx) & 1).astype(np.float32)
+                            keyboard_cond_list.append(multi_hot)
+                        else:
+                            # If already 2D, pad to num_bits if necessary
+                            k_data = action_data.astype(np.float32)
+                            if k_data.ndim == 2 and k_data.shape[-1] < num_bits:
+                                padding = np.zeros((k_data.shape[0], num_bits - k_data.shape[-1]), dtype=np.float32)
+                                k_data = np.concatenate([k_data, padding], axis=-1)
+                            keyboard_cond_list.append(k_data)
             if keyboard_cond_list:
                 features["keyboard_cond"] = keyboard_cond_list
             if mouse_cond_list:

fastvideo/pipelines/preprocess/preprocess_hunyuan_overfit.py

@@ -1,206 +0,0 @@
-# SPDX-License-Identifier: Apache-2.0
-"""Preprocess HunyuanVideo overfit data into parquet format.
-
-Encodes videos with HunyuanVideo VAE and captions with dual text
-encoders (LLaMA + CLIP) into the t2v parquet schema expected by
-the training framework.
-
-Usage:
-    CUDA_VISIBLE_DEVICES=0 python scripts/preprocess_hunyuan_overfit.py
-"""
-
-import glob
-import json
-import os
-
-import cv2
-import numpy as np
-import pyarrow as pa
-import pyarrow.parquet as pq
-import torch
-from safetensors.torch import load_file as safetensors_load_file
-
-from fastvideo.configs.models.vaes import HunyuanVAEConfig
-from fastvideo.configs.pipelines.hunyuan import (
-    clip_preprocess_text,
-    clip_postprocess_text,
-    llama_preprocess_text,
-    llama_postprocess_text,
-)
-from fastvideo.dataset.dataloader.schema import pyarrow_schema_t2v
-from fastvideo.models.vaes.hunyuanvae import AutoencoderKLHunyuanVideo
-from fastvideo.utils import maybe_download_model
-
-# --- Config ---
-NUM_FRAMES = 77  # 4k+1 for temporal compression ratio 4
-MAX_HEIGHT = 480
-MAX_WIDTH = 832
-TRAIN_FPS = 16.0
-
-DATA_DIR = "data/hunyuan_overfit"
-OUTPUT_DIR = "data/hunyuan_overfit_preprocessed"
-
-
-def load_video(path: str, num_frames: int) -> torch.Tensor:
-    """Load video as [1, C, T, H, W] in [-1, 1]."""
-    cap = cv2.VideoCapture(path)
-    frames: list[np.ndarray] = []
-    while len(frames) < num_frames:
-        ret, frame = cap.read()
-        if not ret:
-            break
-        frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
-        frames.append(frame)
-    cap.release()
-
-    if len(frames) < num_frames:
-        # Repeat last frame to fill
-        while len(frames) < num_frames:
-            frames.append(frames[-1])
-
-    frames = frames[:num_frames]
-    video = np.stack(frames, axis=0)
-    video = torch.from_numpy(video).float()
-    video = video / 127.5 - 1.0  # [0,255] -> [-1,1]
-    video = video.permute(3, 0, 1, 2).unsqueeze(0)  # [1,C,T,H,W]
-    return video
-
-
-def main() -> None:
-    device = torch.device("cuda:0")
-    model_path = maybe_download_model("hunyuanvideo-community/HunyuanVideo")
-    vae_path = os.path.join(model_path, "vae")
-
-    os.makedirs(OUTPUT_DIR, exist_ok=True)
-
-    # Load captions
-    with open(os.path.join(DATA_DIR, "videos2caption.json")) as f:
-        caption_data = json.load(f)
-
-    # --- Load VAE ---
-    print("Loading HunyuanVideo VAE...")
-    vae_config = HunyuanVAEConfig()
-    vae_config.load_encoder = True
-    vae_config.load_decoder = False
-    vae = AutoencoderKLHunyuanVideo(vae_config)
-    sf_files = glob.glob(os.path.join(vae_path, "*.safetensors"))
-    loaded = {}
-    for sf_file in sf_files:
-        loaded.update(safetensors_load_file(sf_file))
-    vae.load_state_dict(loaded, strict=False)
-    vae = vae.to(device=device, dtype=torch.float16).eval()
-    vae.use_parallel_tiling = False
-    print(f"VAE loaded ({sum(p.numel() for p in vae.parameters())/1e6:.0f}M)")
-
-    # --- Load text encoders ---
-    print("Loading LLaMA text encoder...")
-    from transformers import AutoTokenizer, CLIPTextModel, LlamaModel
-    from fastvideo.configs.pipelines.hunyuan import (LlamaConfig, CLIPTextConfig)
-
-    llama_cfg = LlamaConfig()
-    llama_tok = AutoTokenizer.from_pretrained(os.path.join(model_path, "tokenizer"))
-    llama_enc = LlamaModel.from_pretrained(
-        os.path.join(model_path, "text_encoder"),
-        torch_dtype=torch.float16,
-    ).to(device).eval()
-    llama_tok_kwargs = dict(llama_cfg.tokenizer_kwargs)
-
-    print("Loading CLIP text encoder...")
-    clip_cfg = CLIPTextConfig()
-    clip_tok = AutoTokenizer.from_pretrained(os.path.join(model_path, "tokenizer_2"))
-    clip_enc = CLIPTextModel.from_pretrained(
-        os.path.join(model_path, "text_encoder_2"),
-        torch_dtype=torch.float16,
-    ).to(device).eval()
-    clip_tok_kwargs = dict(clip_cfg.tokenizer_kwargs)
-
-    # --- Process each video ---
-    records = []
-    for item in caption_data:
-        video_name = item["path"]
-        caption = item["cap"][0]
-        video_path = os.path.join(DATA_DIR, "videos", video_name)
-
-        print(f"\nProcessing: {video_name}")
-        print(f"  Caption: {caption[:80]}...")
-
-        # Encode video
-        video = load_video(video_path, NUM_FRAMES).to(device=device, dtype=torch.float16)
-        print(f"  Video shape: {video.shape}")
-
-        with torch.no_grad():
-            latent_dist = vae.encode(video)
-            # Cast to fp32 — dataloader hardcodes np.float32
-            latent = latent_dist.mean.squeeze(0).float().cpu()
-        print(f"  Latent shape: {latent.shape}")
-
-        # Encode text with LLaMA
-        llama_text = llama_preprocess_text(caption)
-        with torch.no_grad():
-            llama_inputs = llama_tok(llama_text, **llama_tok_kwargs).to(device)
-            llama_out = llama_enc(**llama_inputs, output_hidden_states=True)
-            llama_embeds = llama_postprocess_text(llama_out).squeeze(0)  # [seq, dim]
-
-        # Encode text with CLIP
-        clip_text = clip_preprocess_text(caption)
-        with torch.no_grad():
-            clip_inputs = clip_tok(clip_text, **clip_tok_kwargs).to(device)
-            clip_out = clip_enc(**clip_inputs)
-            clip_pooled = clip_postprocess_text(clip_out).squeeze(0)  # [dim]
-
-        # Combine: [pooled_clip_row, llama_embeds]
-        llama_dim = llama_embeds.shape[-1]
-        pooled_row = torch.zeros(llama_dim, device=device, dtype=torch.float16)
-        pooled_row[:clip_pooled.shape[-1]] = clip_pooled
-        text_embedding = torch.cat(
-            [pooled_row.unsqueeze(0), llama_embeds],
-            dim=0,
-        ).float().cpu()  # [seq+1, dim]
-        print(f"  Text embedding shape: {text_embedding.shape}")
-
-        record = {
-            "id": video_name,
-            "vae_latent_bytes": latent.numpy().tobytes(),
-            "vae_latent_shape": list(latent.shape),
-            "vae_latent_dtype": str(latent.dtype).replace("torch.", ""),
-            "text_embedding_bytes": text_embedding.numpy().tobytes(),
-            "text_embedding_shape": list(text_embedding.shape),
-            "text_embedding_dtype": str(text_embedding.dtype).replace("torch.", ""),
-            "file_name": video_name,
-            "caption": caption,
-            "media_type": "video",
-            "width": MAX_WIDTH,
-            "height": MAX_HEIGHT,
-            "num_frames": NUM_FRAMES,
-            "duration_sec": NUM_FRAMES / TRAIN_FPS,
-            "fps": TRAIN_FPS,
-        }
-        records.append(record)
-
-    # Clean up encoders
-    del llama_enc, llama_tok, clip_enc, clip_tok, vae
-
-    # Write parquet
-    table = pa.table(
-        {k: [r[k] for r in records]
-         for k in records[0]},
-        schema=pyarrow_schema_t2v,
-    )
-    output_path = os.path.join(OUTPUT_DIR, "data_00000.parquet")
-    pq.write_table(table, output_path)
-    print(f"\nWrote {len(records)} records to {output_path}")
-
-    # Write validation prompts for callback
-    # Wrap in "data" key — ValidationDataset expects field="data"
-    # Use "caption" field — ValidationDataset aliases it to "prompt"
-    val_prompts = {"data": [{"caption": item["cap"][0]} for item in caption_data]}
-    val_path = os.path.join(OUTPUT_DIR, "validation_prompts.json")
-    with open(val_path, "w") as f:
-        json.dump(val_prompts, f, indent=2)
-    print(f"Wrote validation prompts to {val_path}")
-
-    print("\nDone! Use data_path: " + OUTPUT_DIR + " in training config.")
-
-
-if __name__ == "__main__":
-    main()

fastvideo/pipelines/preprocess/preprocess_pipeline_i2v.py

@@ -31,12 +31,11 @@ class PreprocessPipeline_I2V(BasePreprocessPipeline):
                            tokenizers=[self.get_module("tokenizer")],
                        ))
 
-        if (self.get_module("image_encoder") is not None and self.get_module("image_processor") is not None):
-            self.add_stage(stage_name="image_encoding_stage",
-                           stage=ImageEncodingStage(
-                               image_encoder=self.get_module("image_encoder"),
-                               image_processor=self.get_module("image_processor"),
-                           ))
+        self.add_stage(stage_name="image_encoding_stage",
+                       stage=ImageEncodingStage(
+                           image_encoder=self.get_module("image_encoder"),
+                           image_processor=self.get_module("image_processor"),
+                       ))
 
     def get_pyarrow_schema(self):
         """Return the PyArrow schema for I2V pipeline."""
@@ -45,34 +44,35 @@ class PreprocessPipeline_I2V(BasePreprocessPipeline):
     def get_extra_features(self, valid_data: dict[str, Any], fastvideo_args: FastVideoArgs) -> dict[str, Any]:
 
         # TODO(will): move these to cpu at some point
+        self.get_module("image_encoder").to(get_local_torch_device())
         self.get_module("vae").to(get_local_torch_device())
 
         features = {}
+        """Get CLIP features from the first frame of each video."""
         first_frame = valid_data["pixel_values"][:, :, 0, :, :].permute(0, 2, 3, 1)  # (B, C, T, H, W) -> (B, H, W, C)
         _, _, num_frames, height, width = valid_data["pixel_values"].shape
+        # latent_height = height // self.get_module(
+        #     "vae").spatial_compression_ratio
+        # latent_width = width // self.get_module("vae").spatial_compression_ratio
 
-        # Get CLIP features if image_encoder is available
-        if (self.get_module("image_encoder") is not None and self.get_module("image_processor") is not None):
-            self.get_module("image_encoder").to(get_local_torch_device())
+        processed_images = []
+        # Frame has values between -1 and 1
+        for frame in first_frame:
+            frame = (frame + 1) * 127.5
+            frame_pil = Image.fromarray(frame.cpu().numpy().astype(np.uint8))
+            processed_img = self.get_module("image_processor")(images=frame_pil, return_tensors="pt")
+            processed_images.append(processed_img)
 
-            processed_images = []
-            for frame in first_frame:
-                frame = (frame + 1) * 127.5
-                frame_pil = Image.fromarray(frame.cpu().numpy().astype(np.uint8))
-                processed_img = self.get_module("image_processor")(images=frame_pil, return_tensors="pt")
-                processed_images.append(processed_img)
+        # Get CLIP features
+        pixel_values = torch.cat([img['pixel_values'] for img in processed_images], dim=0).to(get_local_torch_device())
+        with torch.no_grad():
+            image_inputs = {'pixel_values': pixel_values}
+            with set_forward_context(current_timestep=0, attn_metadata=None):
+                clip_features = self.get_module("image_encoder")(**image_inputs)
+            clip_features = clip_features.last_hidden_state
 
-            pixel_values = torch.cat([img['pixel_values'] for img in processed_images],
-                                     dim=0).to(get_local_torch_device())
-            with torch.no_grad():
-                image_inputs = {'pixel_values': pixel_values}
-                with set_forward_context(current_timestep=0, attn_metadata=None):
-                    clip_features = self.get_module("image_encoder")(**image_inputs)
-                clip_features = clip_features.last_hidden_state
-
-            features["clip_feature"] = clip_features
-
-        # Get VAE features from the first frame of each video
+        features["clip_feature"] = clip_features
+        """Get VAE features from the first frame of each video"""
         video_conditions = []
         for frame in first_frame:
             processed_img = frame.to(device="cpu", dtype=torch.float32)
@@ -105,6 +105,25 @@ class PreprocessPipeline_I2V(BasePreprocessPipeline):
         else:
             latent_condition = latent_condition * self.get_module("vae").scaling_factor
 
+        # mask_lat_size = torch.ones(batch_size, 1, num_frames, latent_height,
+        #                            latent_width)
+        # mask_lat_size[:, :, list(range(1, num_frames))] = 0
+        # first_frame_mask = mask_lat_size[:, :, 0:1]
+        # first_frame_mask = torch.repeat_interleave(
+        #     first_frame_mask,
+        #     dim=2,
+        #     repeats=self.get_module("vae").temporal_compression_ratio)
+        # mask_lat_size = torch.concat(
+        #     [first_frame_mask, mask_lat_size[:, :, 1:, :]], dim=2)
+        # mask_lat_size = mask_lat_size.view(
+        #     batch_size, -1,
+        #     self.get_module("vae").temporal_compression_ratio, latent_height,
+        #     latent_width)
+        # mask_lat_size = mask_lat_size.transpose(1, 2)
+        # mask_lat_size = mask_lat_size.to(latent_condition.device)
+
+        # image_latent = torch.concat([mask_lat_size, latent_condition], dim=1)
+
         features["first_frame_latent"] = latent_condition
 
         return features

fastvideo/pipelines/stages/input_validation.py

@@ -178,6 +178,8 @@ class InputValidationStage(PipelineStage):
             if batch.keyboard_cond.dim() != 3:
                 raise ValueError(f"keyboard_cond must have 3 dimensions (B, T, K), but got {batch.keyboard_cond.dim()}")
             keyboard_dim = batch.keyboard_cond.shape[-1]
+            if keyboard_dim not in {2, 3, 4, 6, 7}:
+                raise ValueError(f"keyboard_cond last dimension must be 2, 3, 4, 6, or 7, but got {keyboard_dim}")
             logger.info("Action control: keyboard_cond validated - shape %s (dim=%d)", batch.keyboard_cond.shape,
                         keyboard_dim)
 

fastvideo/registry.py

@@ -348,18 +348,6 @@ def _register_configs() -> None:
         model_detectors=[lambda path: ("lingbotworld" in path.lower() or "lingbot-world" in path.lower())],
     )
 
-    # Kandinsky5 Lite T2V
-    register_configs(
-        sampling_param_cls=None,
-        pipeline_config_cls=PipelineConfig,
-        hf_model_paths=[
-            "kandinskylab/Kandinsky-5.0-T2V-Lite-sft-5s-Diffusers",
-        ],
-        model_detectors=[
-            lambda path: any(token in path.lower() for token in ("kandinsky5", "kandinsky-5")),
-        ],
-    )
-
     # LongCat
     register_configs(
         sampling_param_cls=None,

fastvideo/tests/encoders/test_clip_encoder.py

@@ -21,7 +21,9 @@ os.environ["MASTER_ADDR"] = "localhost"
 os.environ["MASTER_PORT"] = "29503"
 
 BASE_MODEL_PATH = "hunyuanvideo-community/HunyuanVideo"
-MODEL_PATH = maybe_download_model(BASE_MODEL_PATH, local_dir=os.path.join("data", BASE_MODEL_PATH))
+MODEL_PATH = maybe_download_model(BASE_MODEL_PATH,
+                                  local_dir=os.path.join("data", BASE_MODEL_PATH) # store in the large /workspace disk on Runpod
+                                  )
 TEXT_ENCODER_PATH = os.path.join(MODEL_PATH, "text_encoder_2")
 TOKENIZER_PATH = os.path.join(MODEL_PATH, "tokenizer_2")
 
@@ -32,12 +34,13 @@ def test_clip_encoder():
     Tests compatibility between two different implementations for loading text encoders:
     1. load_text_encoder from fastvideo.models.hunyuan.text_encoder
     2. TextEncoderLoader from fastvideo.models.loader
-
+    
     The test verifies that both implementations:
     - Load models with the same weights and parameters
     - Produce nearly identical outputs for the same input prompts
     """
-    args = FastVideoArgs(model_path="openai/clip-vit-large-patch14", pipeline_config=HunyuanConfig())
+    args = FastVideoArgs(model_path="openai/clip-vit-large-patch14",
+                         pipeline_config=HunyuanConfig())
     device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
     logger.info("Loading models from %s", args.model_path)
@@ -52,7 +55,6 @@ def test_clip_encoder():
     model1 = CLIPTextModel.from_pretrained(TEXT_ENCODER_PATH).to(torch.float16).to(device).eval()
 
     from fastvideo.models.loader.component_loader import TextEncoderLoader
-
     loader = TextEncoderLoader()
     model2 = loader.load(TEXT_ENCODER_PATH, args)
 
@@ -88,7 +90,10 @@ def test_clip_encoder():
     tokenizer = AutoTokenizer.from_pretrained(TOKENIZER_PATH)
 
     # Test with some sample prompts
-    prompts = ["a photo of a cat", "a beautiful landscape with mountains", "an astronaut riding a horse on the moon"]
+    prompts = [
+        "a photo of a cat", "a beautiful landscape with mountains",
+        "an astronaut riding a horse on the moon"
+    ]
 
     logger.info("Testing CLIP text encoder with sample prompts")
 
@@ -97,11 +102,14 @@ def test_clip_encoder():
             logger.info("Testing prompt: '%s'", prompt)
 
             # Tokenize the prompt
-            tokens = tokenizer(prompt, padding="max_length", max_length=77, truncation=True, return_tensors="pt").to(
-                device
-            )
+            tokens = tokenizer(prompt,
+                               padding="max_length",
+                               max_length=77,
+                               truncation=True,
+                               return_tensors="pt").to(device)
             # Get embeddings from our implementation
-            outputs1 = model1(input_ids=tokens.input_ids, output_hidden_states=True)
+            outputs1 = model1(input_ids=tokens.input_ids,
+                              output_hidden_states=True)
 
             logger.info("Testing model2")
             print("--------------------------------")
@@ -110,25 +118,24 @@ def test_clip_encoder():
                 outputs2 = model2(
                     input_ids=tokens.input_ids,
                     # attention_mask=tokens.attention_mask,
-                    output_hidden_states=True,
-                )
+                    output_hidden_states=True)
 
             # Compare last hidden states
-            last_hidden_state1 = outputs1.last_hidden_state[tokens.attention_mask == 1]
-            last_hidden_state2 = outputs2.last_hidden_state[tokens.attention_mask == 1]
+            last_hidden_state1 = outputs1.last_hidden_state[
+                tokens.attention_mask == 1]
+            last_hidden_state2 = outputs2.last_hidden_state[
+                tokens.attention_mask == 1]
             # print("last_hidden_state1", last_hidden_state1)
             # print("last_hidden_state2", last_hidden_state2)
 
-            assert last_hidden_state1.shape == last_hidden_state2.shape, (
+            assert last_hidden_state1.shape == last_hidden_state2.shape, \
                 f"Hidden state shapes don't match: {last_hidden_state1.shape} vs {last_hidden_state2.shape}"
-            )
             # Compare pooler outputs
             pooler_output1 = outputs1.pooler_output
             pooler_output2 = outputs2.pooler_output
 
-            assert pooler_output1.shape == pooler_output2.shape, (
+            assert pooler_output1.shape == pooler_output2.shape, \
                 f"Pooler output shapes don't match: {pooler_output1.shape} vs {pooler_output2.shape}"
-            )
 
             assert_close(pooler_output1, pooler_output2, atol=1e-2, rtol=1e-3)
             assert_close(last_hidden_state1, last_hidden_state2, atol=1e-2, rtol=1e-3)

fastvideo/tests/encoders/test_llama_encoder.py

@@ -15,14 +15,15 @@ from fastvideo.utils import maybe_download_model
 from fastvideo.configs.models.encoders import LlamaConfig
 from torch.distributed.tensor import DTensor
 from torch.testing import assert_close
-
 logger = init_logger(__name__)
 
 os.environ["MASTER_ADDR"] = "localhost"
 os.environ["MASTER_PORT"] = "29503"
 
 BASE_MODEL_PATH = "hunyuanvideo-community/HunyuanVideo"
-MODEL_PATH = maybe_download_model(BASE_MODEL_PATH, local_dir=os.path.join("data", BASE_MODEL_PATH))
+MODEL_PATH = maybe_download_model(BASE_MODEL_PATH,
+                                  local_dir=os.path.join("data", BASE_MODEL_PATH) # store in the large /workspace disk on Runpod
+                                  )
 TEXT_ENCODER_PATH = os.path.join(MODEL_PATH, "text_encoder")
 TOKENIZER_PATH = os.path.join(MODEL_PATH, "tokenizer")
 
@@ -33,12 +34,13 @@ def test_llama_encoder():
     Tests compatibility between two different implementations for loading text encoders:
     1. load_text_encoder from fastvideo.models.hunyuan.text_encoder
     2. TextEncoderLoader from fastvideo.models.loader
-
+    
     The test verifies that both implementations:
     - Load models with the same weights and parameters
     - Produce nearly identical outputs for the same input prompts
     """
-    args = FastVideoArgs(model_path="meta-llama/Llama-2-7b-hf", pipeline_config=HunyuanConfig())
+    args = FastVideoArgs(model_path="meta-llama/Llama-2-7b-hf",
+                         pipeline_config=HunyuanConfig())
 
     device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 
@@ -66,15 +68,15 @@ def test_llama_encoder():
     logger.info("Model1 has %d parameters", len(params1))
     logger.info("Model2 has %d parameters", len(params2))
 
+
     # check if embed_tokens are the same
     device = model1.embed_tokens.weight.device
-    assert torch.allclose(
-        model1.embed_tokens.weight,
-        model2.embed_tokens.weight.to_local().to(device)
-        if isinstance(model2.embed_tokens.weight, DTensor)
-        else model2.embed_tokens.weight.to(device),
-    )
-    weights = ["layers.{}.input_layernorm.weight", "layers.{}.post_attention_layernorm.weight"]
+    assert torch.allclose(model1.embed_tokens.weight,
+                          model2.embed_tokens.weight.to_local().to(device) if isinstance(model2.embed_tokens.weight, DTensor) else model2.embed_tokens.weight.to(device))
+    weights = [
+        "layers.{}.input_layernorm.weight",
+        "layers.{}.post_attention_layernorm.weight"
+    ]
     for layer_idx in range(hf_config.num_hidden_layers):
         for w in weights:
             name1 = w.format(layer_idx)
@@ -119,29 +121,33 @@ def test_llama_encoder():
             logger.info("Testing prompt: '%s'", prompt)
 
             # Tokenize the prompt
-            tokens = tokenizer(prompt, padding="max_length", max_length=512, truncation=True, return_tensors="pt").to(
-                device
-            )
+            tokens = tokenizer(prompt,
+                               padding="max_length",
+                               max_length=512,
+                               truncation=True,
+                               return_tensors="pt").to(device)
 
             # Get outputs from our implementation
             # filter out padding input_ids
             # tokens.input_ids = tokens.input_ids[tokens.attention_mask==1]
             # tokens.attention_mask = tokens.attention_mask[tokens.attention_mask==1]
-            outputs1 = model1(input_ids=tokens.input_ids, output_hidden_states=True)
+            outputs1 = model1(input_ids=tokens.input_ids,
+                              output_hidden_states=True)
             print("--------------------------------")
             logger.info("Testing model2")
 
             # Get outputs from HuggingFace implementation
             with set_forward_context(current_timestep=0, attn_metadata=None):
-                outputs2 = model2(
-                    input_ids=tokens.input_ids, attention_mask=tokens.attention_mask, output_hidden_states=True
-                )
+                outputs2 = model2(input_ids=tokens.input_ids,
+                                  attention_mask=tokens.attention_mask,
+                                  output_hidden_states=True)
 
             # Compare last hidden states
-            last_hidden_state1 = outputs1.last_hidden_state[tokens.attention_mask == 1]
-            last_hidden_state2 = outputs2.last_hidden_state[tokens.attention_mask == 1]
+            last_hidden_state1 = outputs1.last_hidden_state[
+                tokens.attention_mask == 1]
+            last_hidden_state2 = outputs2.last_hidden_state[
+                tokens.attention_mask == 1]
 
-            assert last_hidden_state1.shape == last_hidden_state2.shape, (
+            assert last_hidden_state1.shape == last_hidden_state2.shape, \
                 f"Hidden state shapes don't match: {last_hidden_state1.shape} vs {last_hidden_state2.shape}"
-            )
             assert_close(last_hidden_state1, last_hidden_state2, atol=1e-1, rtol=1e-4)

fastvideo/tests/modal/ssim_test.py

@@ -1,11 +1,7 @@
 import ast
-import datetime
 import glob
 import os
 import re
-import shutil
-import subprocess
-import sys
 from dataclasses import dataclass
 from typing import Any
 
@@ -14,7 +10,7 @@ import modal
 app = modal.App()
 
 model_vol = modal.Volume.from_name("hf-model-weights")
-image_version = os.getenv("IMAGE_VERSION", "latest")
+image_version = os.getenv("IMAGE_VERSION")
 image_tag = f"ghcr.io/hao-ai-lab/fastvideo/fastvideo-dev:{image_version}"
 print(f"Using image: {image_tag}")
 
@@ -28,29 +24,30 @@ image = (
         "libssl-dev",
         "ffmpeg",
     )
-    .run_commands("curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --default-toolchain stable")
-    .run_commands("echo 'source ~/.cargo/env' >> ~/.bashrc")
-    .env(
-        {
-            "PATH": "/root/.cargo/bin:$PATH",
-            "BUILDKITE_REPO": os.environ.get("BUILDKITE_REPO", ""),
-            "BUILDKITE_COMMIT": os.environ.get("BUILDKITE_COMMIT", ""),
-            "BUILDKITE_PULL_REQUEST": os.environ.get("BUILDKITE_PULL_REQUEST", ""),
-            "IMAGE_VERSION": os.environ.get("IMAGE_VERSION", ""),
-        }
+    .run_commands(
+        "curl --proto '=https' --tlsv1.2 -sSf "
+        "https://sh.rustup.rs | sh -s -- -y --default-toolchain stable"
     )
+    .run_commands("echo 'source ~/.cargo/env' >> ~/.bashrc")
+    .env({
+        "PATH": "/root/.cargo/bin:$PATH",
+        "BUILDKITE_REPO": os.environ.get("BUILDKITE_REPO", ""),
+        "BUILDKITE_COMMIT": os.environ.get("BUILDKITE_COMMIT", ""),
+        "BUILDKITE_PULL_REQUEST": os.environ.get(
+            "BUILDKITE_PULL_REQUEST", ""
+        ),
+        "IMAGE_VERSION": os.environ.get("IMAGE_VERSION", ""),
+    })
 )
 
 SSIM_NUM_GPUS = 4
 SSIM_TERMINATE_TIMEOUT_S = 30
-HF_TOKEN_ENV_KEYS = ("HF_API_KEY", "HUGGINGFACE_HUB_TOKEN", "HF_TOKEN")
-RAW_GENERATED_VOLUME_ROOT = "ssim_generated_videos"
-DEFAULT_OUTPUT_QUALITY_TIER = "default"
-FULL_OUTPUT_QUALITY_TIER = "full_quality"
-MODAL_DEVICE_REFERENCE_FOLDER = "L40S_reference_videos"
 SSIM_COMMON_KWARGS = dict(
     image=image,
     timeout=5400,
+    secrets=[modal.Secret.from_dict(
+        {"HF_API_KEY": os.environ.get("HF_API_KEY", "")}
+    )],
     volumes={"/root/data": model_vol},
 )
 
@@ -108,194 +105,6 @@ class _PartitionResult:
     exit_code: int
 
 
-def _split_csv_values(csv_values: str) -> set[str]:
-    return {value.strip() for value in csv_values.split(",") if value.strip()}
-
-
-def _run_git_command(args: list[str]) -> str:
-    try:
-        result = subprocess.run(
-            ["git", *args],
-            check=True,
-            capture_output=True,
-            text=True,
-        )
-    except subprocess.CalledProcessError as error:
-        stderr = (error.stderr or "").strip()
-        raise RuntimeError(f"Failed to run git {' '.join(args)}: {stderr or error}") from error
-    return result.stdout.strip()
-
-
-def _normalize_git_repo_url(git_repo: str) -> str:
-    if git_repo.startswith("git@github.com:"):
-        return "https://github.com/" + git_repo[len("git@github.com:") :]
-    if git_repo.startswith("ssh://git@github.com/"):
-        return "https://github.com/" + git_repo[len("ssh://git@github.com/") :]
-    return git_repo
-
-
-def _resolve_git_repo(git_repo: str) -> str:
-    if git_repo.strip():
-        return _normalize_git_repo_url(git_repo.strip())
-
-    env_repo = os.environ.get("BUILDKITE_REPO", "").strip()
-    if env_repo:
-        return _normalize_git_repo_url(env_repo)
-
-    discovered_repo = _run_git_command(["config", "--get", "remote.origin.url"])
-    if discovered_repo:
-        return _normalize_git_repo_url(discovered_repo)
-
-    raise RuntimeError("Could not resolve git repo URL. Pass --git-repo or set BUILDKITE_REPO.")
-
-
-def _resolve_git_commit(git_commit: str) -> str:
-    if git_commit.strip():
-        return git_commit.strip()
-
-    env_commit = os.environ.get("BUILDKITE_COMMIT", "").strip()
-    if env_commit:
-        return env_commit
-
-    discovered_commit = _run_git_command(["rev-parse", "HEAD"])
-    if discovered_commit:
-        return discovered_commit
-
-    raise RuntimeError("Could not resolve git commit. Pass --git-commit or set BUILDKITE_COMMIT.")
-
-
-def _resolve_pull_request(pr_number: str) -> str:
-    if pr_number.strip():
-        return pr_number.strip()
-    env_pr = os.environ.get("BUILDKITE_PULL_REQUEST", "").strip()
-    if env_pr:
-        return env_pr
-    return "false"
-
-
-def _resolve_hf_api_key(
-    hf_api_key: str,
-) -> str:
-    if hf_api_key.strip():
-        return hf_api_key.strip()
-
-    for key in HF_TOKEN_ENV_KEYS:
-        value = os.environ.get(key, "").strip()
-        if value:
-            return value
-
-    raise RuntimeError(
-        "Hugging Face token is required. Set HF_API_KEY, HUGGINGFACE_HUB_TOKEN, or HF_TOKEN; or pass --hf-api-key."
-    )
-
-
-def _sanitize_path_fragment(value: str) -> str:
-    cleaned = re.sub(r"[^A-Za-z0-9_.-]+", "_", value.strip())
-    return cleaned.strip("._-")
-
-
-def _resolve_generated_volume_subdir(
-    requested_subdir: str,
-    git_commit: str,
-) -> str:
-    if requested_subdir.strip():
-        normalized = requested_subdir.strip().strip("/")
-        if not normalized:
-            raise RuntimeError("generated_volume_subdir must not be empty.")
-        return normalized
-
-    short_commit = _sanitize_path_fragment(git_commit[:12]) or "unknown"
-    timestamp = datetime.datetime.utcnow().strftime("%Y%m%d_%H%M%S")
-    return f"{timestamp}_{short_commit}"
-
-
-def _resolve_output_quality_tier(ssim_full_quality: bool) -> str:
-    if ssim_full_quality:
-        return FULL_OUTPUT_QUALITY_TIER
-    return DEFAULT_OUTPUT_QUALITY_TIER
-
-
-def _build_generated_volume_relative_path(
-    *,
-    generated_volume_subdir: str,
-    quality_tier: str,
-) -> str:
-    return os.path.join(
-        RAW_GENERATED_VOLUME_ROOT,
-        quality_tier,
-        generated_volume_subdir,
-        "generated_videos",
-    )
-
-
-def _build_local_generated_download_dir(quality_tier: str) -> str:
-    return os.path.join(".", "generated_videos_modal", quality_tier)
-
-
-def _print_local_reference_copy_command(quality_tier: str) -> None:
-    generated_dir = os.path.join(
-        _build_local_generated_download_dir(quality_tier),
-        MODAL_DEVICE_REFERENCE_FOLDER,
-    )
-    print(
-        "To update local references from the downloaded Modal outputs, run:\n"
-        "  python fastvideo/tests/ssim/reference_videos_cli.py copy-local "
-        f"--quality-tier {quality_tier} "
-        f"--generated-dir {generated_dir} "
-        f"--device-folder {MODAL_DEVICE_REFERENCE_FOLDER}"
-    )
-
-
-def _count_video_files(root: str) -> int:
-    count = 0
-    for current_root, _, filenames in os.walk(root):
-        for filename in filenames:
-            if filename.lower().endswith((".mp4", ".avi", ".mov", ".mkv", ".webm", ".flv")):
-                count += 1
-    return count
-
-
-def _sync_generated_videos_to_volume(
-    repo_root: str,
-    generated_volume_subdir: str,
-    quality_tier: str,
-) -> str | None:
-    generated_root = os.path.join(
-        repo_root,
-        "fastvideo",
-        "tests",
-        "ssim",
-        "generated_videos",
-        quality_tier,
-    )
-    if not os.path.isdir(generated_root):
-        print(
-            f"No generated_videos directory found for quality tier {quality_tier}; skipping raw generated video export."
-        )
-        return None
-
-    relative_dst = _build_generated_volume_relative_path(
-        generated_volume_subdir=generated_volume_subdir,
-        quality_tier=quality_tier,
-    )
-    absolute_dst = os.path.join("/root/data", relative_dst)
-    if os.path.exists(absolute_dst):
-        shutil.rmtree(absolute_dst)
-    os.makedirs(os.path.dirname(absolute_dst), exist_ok=True)
-    shutil.copytree(generated_root, absolute_dst)
-    model_vol.commit()
-
-    num_videos = _count_video_files(absolute_dst)
-    print(f"Raw generated videos exported to Modal volume path: {relative_dst} ({num_videos} video files).")
-    print(
-        "Download command:\n"
-        f"  modal volume get hf-model-weights {relative_dst} "
-        f"{_build_local_generated_download_dir(quality_tier)}"
-    )
-    _print_local_reference_copy_command(quality_tier)
-    return relative_dst
-
-
 def _extract_required_gpus(filepath: str) -> int:
     """Read REQUIRED_GPUS from a test file. Defaults to 1."""
     with open(filepath, encoding="utf-8") as file:
@@ -317,7 +126,10 @@ def _extract_model_ids(filepath: str) -> list[str]:
         value_node = None
 
         if isinstance(node, ast.Assign):
-            target_names = [target.id for target in node.targets if isinstance(target, ast.Name)]
+            target_names = [
+                target.id for target in node.targets
+                if isinstance(target, ast.Name)
+            ]
             value_node = node.value
         elif isinstance(node, ast.AnnAssign):
             if isinstance(node.target, ast.Name):
@@ -326,7 +138,8 @@ def _extract_model_ids(filepath: str) -> list[str]:
 
         if not target_names or value_node is None:
             continue
-        if not any(name.endswith("MODEL_TO_PARAMS") for name in target_names):
+        if not any(name.endswith("MODEL_TO_PARAMS")
+                   for name in target_names):
             continue
         if not isinstance(value_node, ast.Dict):
             continue
@@ -346,70 +159,35 @@ def _extract_model_ids(filepath: str) -> list[str]:
     return unique_model_ids
 
 
-def _discover_ssim_tasks(
-    ssim_dir: str,
-    *,
-    selected_test_files: set[str] | None = None,
-    selected_model_ids: set[str] | None = None,
-) -> list[SSIMTask]:
+def _discover_ssim_tasks(ssim_dir: str) -> list[SSIMTask]:
     tasks = []
     task_id = 0
-    selected_test_files = selected_test_files or set()
-    selected_test_file_names = {os.path.basename(test_file) for test_file in selected_test_files}
-    selected_model_ids = selected_model_ids or set()
-
     test_files = sorted(glob.glob(os.path.join(ssim_dir, "test_*.py")))
-    matched_test_files = set()
-    matched_model_ids = set()
     for filepath in test_files:
-        file_name = os.path.basename(filepath)
-        rel_path = f"./fastvideo/tests/ssim/{file_name}"
-        if selected_test_files and (
-            filepath not in selected_test_files
-            and rel_path not in selected_test_files
-            and file_name not in selected_test_file_names
-        ):
-            continue
-
-        matched_test_files.add(file_name)
         required_gpus = _extract_required_gpus(filepath)
         if required_gpus < 1 or required_gpus > SSIM_NUM_GPUS:
-            raise ValueError(f"{filepath} requires {required_gpus} GPUs, but scheduler supports up to {SSIM_NUM_GPUS}.")
+            raise ValueError(
+                f"{filepath} requires {required_gpus} GPUs, "
+                f"but scheduler supports up to {SSIM_NUM_GPUS}."
+            )
+        rel_path = f"./fastvideo/tests/ssim/{os.path.basename(filepath)}"
         model_ids = _extract_model_ids(filepath)
         if model_ids:
             for model_id in model_ids:
-                if selected_model_ids and model_id not in selected_model_ids:
-                    continue
-                matched_model_ids.add(model_id)
-                tasks.append(
-                    SSIMTask(
-                        task_id=task_id,
-                        test_file=rel_path,
-                        required_gpus=required_gpus,
-                        model_id=model_id,
-                    )
-                )
-                task_id += 1
-        else:
-            if selected_model_ids:
-                continue
-            tasks.append(
-                SSIMTask(
+                tasks.append(SSIMTask(
                     task_id=task_id,
                     test_file=rel_path,
                     required_gpus=required_gpus,
-                )
-            )
+                    model_id=model_id,
+                ))
+                task_id += 1
+        else:
+            tasks.append(SSIMTask(
+                task_id=task_id,
+                test_file=rel_path,
+                required_gpus=required_gpus,
+            ))
             task_id += 1
-
-    if selected_test_files:
-        unmatched_files = sorted(selected_test_file_names - matched_test_files)
-        if unmatched_files:
-            raise RuntimeError("Requested SSIM test file(s) not found: " + ", ".join(unmatched_files))
-    if selected_model_ids:
-        unmatched_model_ids = sorted(selected_model_ids - matched_model_ids)
-        if unmatched_model_ids:
-            raise RuntimeError("Requested SSIM model_id(s) not found: " + ", ".join(unmatched_model_ids))
     return sorted(tasks, key=lambda task: task.sort_key)
 
 
@@ -429,24 +207,26 @@ def _build_checkout_command(git_commit: str, pr_number: str | None) -> str:
         try:
             pr_id = int(pr_number)
         except ValueError as error:
-            raise RuntimeError(f"Invalid BUILDKITE_PULL_REQUEST value: {pr_number}") from error
-        return f"git fetch --prune origin refs/pull/{pr_id}/head && git checkout FETCH_HEAD"
+            raise RuntimeError(
+                f"Invalid BUILDKITE_PULL_REQUEST value: {pr_number}"
+            ) from error
+        return (
+            "git fetch --prune origin "
+            f"refs/pull/{pr_id}/head && "
+            "git checkout FETCH_HEAD"
+        )
     return f"git checkout {shlex.quote(git_commit)}"
 
 
-def _prepare_ssim_workspace(
-    *,
-    git_repo: str,
-    git_commit: str,
-    pr_number: str,
-    hf_api_key: str,
-    selected_test_files: set[str] | None = None,
-    selected_model_ids: set[str] | None = None,
-) -> tuple[str, list[SSIMTask]]:
+def _prepare_ssim_workspace() -> tuple[str, list[SSIMTask]]:
     import shlex
+    import subprocess
 
-    if not hf_api_key.strip():
-        raise RuntimeError("HF API key is required to prepare SSIM workspace.")
+    git_repo = os.environ.get("BUILDKITE_REPO")
+    git_commit = os.environ.get("BUILDKITE_COMMIT")
+    pr_number = os.environ.get("BUILDKITE_PULL_REQUEST")
+    if not git_repo or not git_commit:
+        raise RuntimeError("BUILDKITE_REPO and BUILDKITE_COMMIT must be set.")
 
     checkout_command = _build_checkout_command(git_commit, pr_number)
     repo_root = "/FastVideo"
@@ -455,14 +235,7 @@ def _prepare_ssim_workspace(
     set -euo pipefail
     source $HOME/.local/bin/env
     source /opt/venv/bin/activate
-    if [ -d {shlex.quote(repo_root)}/.git ]; then
-      cd {shlex.quote(repo_root)}
-      git remote set-url origin {shlex.quote(git_repo)} || true
-      git fetch --prune origin
-    else
-      git clone {shlex.quote(git_repo)} {shlex.quote(repo_root)}
-      cd {shlex.quote(repo_root)}
-    fi
+    cd {shlex.quote(repo_root)}
     {checkout_command}
     git submodule update --init --recursive
     cd fastvideo-kernel
@@ -476,22 +249,17 @@ def _prepare_ssim_workspace(
         ["/bin/bash", "-lc", command],
         capture_output=True,
         text=True,
-        env={
-            **os.environ,
-            "HF_API_KEY": hf_api_key,
-        },
     )
     if result.returncode != 0:
         print(result.stdout)
         print(result.stderr)
-        raise RuntimeError(f"Workspace setup failed with exit code {result.returncode}")
+        raise RuntimeError(
+            "Workspace setup failed with "
+            f"exit code {result.returncode}"
+        )
 
     ssim_dir = os.path.join(repo_root, "fastvideo", "tests", "ssim")
-    tasks = _discover_ssim_tasks(
-        ssim_dir,
-        selected_test_files=selected_test_files,
-        selected_model_ids=selected_model_ids,
-    )
+    tasks = _discover_ssim_tasks(ssim_dir)
     if not tasks:
         raise RuntimeError("No SSIM test files found.")
     return repo_root, tasks
@@ -503,14 +271,22 @@ def _spawn_ssim_task(
     assigned_gpu_ids: list[str],
     log_dir: str,
     task_index: int,
-    pytest_extra_args: list[str],
 ) -> _RunningTask:
     import shlex
+    import subprocess
 
-    safe_test_name = re.sub(r"[^A-Za-z0-9_.-]+", "_", task.test_name)
-    log_path = os.path.join(log_dir, f"{task_index:03d}_{safe_test_name}.log")
-    pytest_command = shlex.join(["pytest", task.test_file, "-vs", *pytest_extra_args])
-    command = f"set -euo pipefail && source $HOME/.local/bin/env && source /opt/venv/bin/activate && {pytest_command}"
+    safe_test_name = re.sub(
+        r"[^A-Za-z0-9_.-]+", "_", task.test_name
+    )
+    log_path = os.path.join(
+        log_dir, f"{task_index:03d}_{safe_test_name}.log"
+    )
+    command = (
+        "set -euo pipefail && "
+        "source $HOME/.local/bin/env && "
+        "source /opt/venv/bin/activate && "
+        f"pytest {shlex.quote(task.test_file)} -vs"
+    )
     env = os.environ.copy()
     env["HF_HOME"] = "/root/data/.cache"
     # MultiprocExecutor returns CUDA tensors through mp pipes (CUDA IPC).
@@ -551,7 +327,9 @@ def _finalize_running_task(
 ) -> None:
     running_task.log_handle.close()
     available_gpu_ids.extend(running_task.gpu_ids)
-    available_gpu_ids.sort(key=lambda gpu_id: gpu_order.get(gpu_id, len(gpu_order)))
+    available_gpu_ids.sort(
+        key=lambda gpu_id: gpu_order.get(gpu_id, len(gpu_order))
+    )
     results[running_task.task.task_id] = _TaskResult(
         task=running_task.task,
         status=status,
@@ -624,30 +402,9 @@ def _get_visible_gpu_ids() -> list[str]:
     return gpu_ids
 
 
-def _build_pytest_extra_args(
-    *,
-    ssim_full_quality: bool,
-    ssim_reference_repo: str,
-    skip_ssim_reference_download: bool,
-    pytest_k: str,
-) -> list[str]:
-    args = []
-    if ssim_full_quality:
-        args.append("--ssim-full-quality")
-    if ssim_reference_repo.strip():
-        args.extend(["--ssim-reference-repo", ssim_reference_repo.strip()])
-    if skip_ssim_reference_download:
-        args.append("--skip-ssim-reference-download")
-    if pytest_k.strip():
-        args.extend(["-k", pytest_k.strip()])
-    return args
-
-
 def _schedule_ssim_tasks(
     repo_root: str,
     tasks: list[SSIMTask],
-    pytest_extra_args: list[str],
-    fail_fast: bool = True,
 ) -> dict[int, _TaskResult]:
     import tempfile
     import time
@@ -655,10 +412,16 @@ def _schedule_ssim_tasks(
     pending_tasks = list(tasks)
     running_tasks = []
     available_gpu_ids = _get_visible_gpu_ids()
-    gpu_order = {gpu_id: index for index, gpu_id in enumerate(available_gpu_ids)}
-    max_required_gpus = max((task.required_gpus for task in pending_tasks), default=0)
+    gpu_order = {
+        gpu_id: index for index, gpu_id in enumerate(available_gpu_ids)
+    }
+    max_required_gpus = max(
+        (task.required_gpus for task in pending_tasks), default=0
+    )
     if max_required_gpus > len(available_gpu_ids):
-        cuda_visible_devices = os.environ.get("CUDA_VISIBLE_DEVICES", "<unset>")
+        cuda_visible_devices = os.environ.get(
+            "CUDA_VISIBLE_DEVICES", "<unset>"
+        )
         raise RuntimeError(
             "SSIM task requires "
             f"{max_required_gpus} GPUs but only "
@@ -680,17 +443,19 @@ def _schedule_ssim_tasks(
                 break
 
             task = pending_tasks.pop(next_task_index)
-            assigned_gpu_ids = available_gpu_ids[: task.required_gpus]
-            del available_gpu_ids[: task.required_gpus]
+            assigned_gpu_ids = available_gpu_ids[:task.required_gpus]
+            del available_gpu_ids[:task.required_gpus]
             running_task = _spawn_ssim_task(
                 task=task,
                 repo_root=repo_root,
                 assigned_gpu_ids=assigned_gpu_ids,
                 log_dir=log_dir,
                 task_index=task.task_id,
-                pytest_extra_args=pytest_extra_args,
             )
-            print(f"Started {task.test_name} on GPUs {','.join(assigned_gpu_ids)}")
+            print(
+                f"Started {task.test_name} on GPUs "
+                f"{','.join(assigned_gpu_ids)}"
+            )
             running_tasks.append(running_task)
 
         completed_tasks = []
@@ -709,8 +474,11 @@ def _schedule_ssim_tasks(
                 gpu_order=gpu_order,
             )
             running_tasks.remove(running_task)
-            print(f"Finished {running_task.task.test_name} with exit code {returncode}")
-            if returncode != 0 and fail_fast and not fail_fast_triggered:
+            print(
+                f"Finished {running_task.task.test_name} with "
+                f"exit code {returncode}"
+            )
+            if returncode != 0 and not fail_fast_triggered:
                 fail_fast_triggered = True
 
         if fail_fast_triggered and running_tasks:
@@ -749,18 +517,22 @@ def _collect_task_summaries(
     for task in tasks:
         result = results[task.task_id]
         log_content = None
-        if result.status == "failed" and result.log_path and os.path.exists(result.log_path):
-            with open(result.log_path, encoding="utf-8") as f:
+        if (
+            result.status == "failed"
+            and result.log_path
+            and os.path.exists(result.log_path)
+        ):
+            with open(
+                result.log_path, encoding="utf-8"
+            ) as f:
                 log_content = f.read()
-        summaries.append(
-            _TaskSummary(
-                test_name=task.test_name,
-                required_gpus=task.required_gpus,
-                status=result.status,
-                returncode=result.returncode,
-                log_content=log_content,
-            )
-        )
+        summaries.append(_TaskSummary(
+            test_name=task.test_name,
+            required_gpus=task.required_gpus,
+            status=result.status,
+            returncode=result.returncode,
+            log_content=log_content,
+        ))
     return summaries
 
 
@@ -778,13 +550,18 @@ def _print_combined_results(
         if result is None:
             continue
         for s in result.task_summaries:
-            label = f"[P{result.partition_index}] {s.test_name}"
+            label = (
+                f"[P{result.partition_index}] "
+                f"{s.test_name}"
+            )
             if s.status == "passed":
                 passed.append(label)
             elif s.status == "failed":
                 failed.append(label)
                 if first_failure is None:
-                    first_failure = (result.partition_index, s)
+                    first_failure = (
+                        result.partition_index, s
+                    )
             elif s.status == "terminated":
                 terminated.append(label)
             elif s.status == "skipped":
@@ -808,8 +585,14 @@ def _print_combined_results(
             status = "cancelled"
         else:
             count = len(result.task_summaries)
-            status = "passed" if result.exit_code == 0 else "failed"
-        print(f"  Partition {i}: {status} ({count} tasks)")
+            status = (
+                "passed" if result.exit_code == 0
+                else "failed"
+            )
+        print(
+            f"  Partition {i}: {status} "
+            f"({count} tasks)"
+        )
     print(f"  passed: {len(passed)}")
     print(f"  failed: {len(failed)}")
     print(f"  terminated: {len(terminated)}")
@@ -822,72 +605,40 @@ def _print_combined_results(
     if skipped:
         print(f"Skipped: {', '.join(skipped)}")
 
-    has_failures = bool(failed or terminated or skipped or any(r is None for r in partition_results))
+    has_failures = bool(
+        failed or terminated or skipped
+        or any(r is None for r in partition_results)
+    )
     return 1 if has_failures else 0
 
 
 @app.function(gpu=f"L40S:{SSIM_NUM_GPUS}", **SSIM_COMMON_KWARGS)
 def run_ssim_partition(
     partition_index: int,
-    num_partitions: int,
-    git_repo: str,
-    git_commit: str,
-    pr_number: str = "false",
-    hf_api_key: str = "",
-    test_files_csv: str = "",
-    model_ids_csv: str = "",
-    ssim_full_quality: bool = False,
-    ssim_reference_repo: str = "",
-    skip_ssim_reference_download: bool = False,
-    pytest_k: str = "",
-    sync_generated_to_volume: bool = False,
-    generated_volume_subdir: str = "",
-    fail_fast: bool = True,
 ) -> _PartitionResult:
-    selected_test_files = _split_csv_values(test_files_csv)
-    selected_model_ids = _split_csv_values(model_ids_csv)
-    repo_root, tasks = _prepare_ssim_workspace(
-        git_repo=git_repo,
-        git_commit=git_commit,
-        pr_number=pr_number,
-        hf_api_key=hf_api_key,
-        selected_test_files=selected_test_files,
-        selected_model_ids=selected_model_ids,
-    )
-    partition = _partition_tasks(tasks, partition_index, num_partitions)
+    repo_root, tasks = _prepare_ssim_workspace()
+    partition = _partition_tasks(tasks, partition_index)
     if not partition:
-        print(f"Partition {partition_index}: no tasks assigned.")
+        print(
+            f"Partition {partition_index}: "
+            "no tasks assigned."
+        )
         return _PartitionResult(
             partition_index=partition_index,
             task_summaries=[],
             exit_code=0,
         )
-    print(f"Partition {partition_index}: running {len(partition)}/{len(tasks)} tasks")
-    pytest_extra_args = _build_pytest_extra_args(
-        ssim_full_quality=ssim_full_quality,
-        ssim_reference_repo=ssim_reference_repo,
-        skip_ssim_reference_download=skip_ssim_reference_download,
-        pytest_k=pytest_k,
+    print(
+        f"Partition {partition_index}: running "
+        f"{len(partition)}/{len(tasks)} tasks"
     )
-    results = _schedule_ssim_tasks(
-        repo_root,
-        partition,
-        pytest_extra_args=pytest_extra_args,
-        fail_fast=fail_fast,
+    results = _schedule_ssim_tasks(repo_root, partition)
+    summaries = _collect_task_summaries(
+        partition, results
+    )
+    has_failures = any(
+        s.status != "passed" for s in summaries
     )
-    summaries = _collect_task_summaries(partition, results)
-    has_failures = any(s.status != "passed" for s in summaries)
-    if sync_generated_to_volume:
-        quality_tier = _resolve_output_quality_tier(ssim_full_quality)
-        resolved_subdir = _resolve_generated_volume_subdir(
-            generated_volume_subdir,
-            git_commit,
-        )
-        _sync_generated_videos_to_volume(
-            repo_root,
-            resolved_subdir,
-            quality_tier,
-        )
     return _PartitionResult(
         partition_index=partition_index,
         task_summaries=summaries,
@@ -895,91 +646,18 @@ def run_ssim_partition(
     )
 
 
-NUM_PARTITIONS = 2
-
-
 @app.local_entrypoint()
-def run_ssim_tests(
-    git_repo: str = "",
-    git_commit: str = "",
-    pr_number: str = "",
-    hf_api_key: str = "",
-    test_files: str = "",
-    model_ids: str = "",
-    full_quality: bool = False,
-    reference_repo: str = "",
-    skip_reference_download: bool = False,
-    pytest_k: str = "",
-    sync_generated_to_volume: bool = False,
-    generated_volume_subdir: str = "",
-    no_fail_fast: bool = False,
-):
-    resolved_git_repo = _resolve_git_repo(git_repo)
-    resolved_git_commit = _resolve_git_commit(git_commit)
-    resolved_pr_number = _resolve_pull_request(pr_number)
-    resolved_hf_api_key = _resolve_hf_api_key(hf_api_key)
+def run_ssim_tests():
+    import sys
 
-    print(f"Running SSIM on repo: {resolved_git_repo}")
-    print(f"Using commit: {resolved_git_commit}")
-    if resolved_pr_number and resolved_pr_number != "false":
-        print(f"Using PR ref: {resolved_pr_number}")
-    if test_files.strip():
-        print(f"Selected test files: {test_files}")
-    if model_ids.strip():
-        print(f"Selected model ids: {model_ids}")
-    if pytest_k.strip():
-        print(f"Using pytest -k filter: {pytest_k}")
-    quality_tier = _resolve_output_quality_tier(full_quality)
-    if sync_generated_to_volume:
-        resolved_subdir = _resolve_generated_volume_subdir(
-            generated_volume_subdir,
-            resolved_git_commit,
-        )
-        print(
-            "Raw generated videos will be saved to Modal volume path: "
-            f"{_build_generated_volume_relative_path(generated_volume_subdir=resolved_subdir, quality_tier=quality_tier)}"
-        )
-    else:
-        resolved_subdir = ""
-
-    common_kwargs = dict(
-        git_repo=resolved_git_repo,
-        git_commit=resolved_git_commit,
-        pr_number=resolved_pr_number,
-        hf_api_key=resolved_hf_api_key,
-        test_files_csv=test_files,
-        model_ids_csv=model_ids,
-        ssim_full_quality=full_quality,
-        ssim_reference_repo=reference_repo,
-        skip_ssim_reference_download=skip_reference_download,
-        pytest_k=pytest_k,
-        sync_generated_to_volume=sync_generated_to_volume,
-        generated_volume_subdir=resolved_subdir,
-        fail_fast=not no_fail_fast,
-    )
-    futures = [
-        run_ssim_partition.spawn(
-            partition_index=i,
-            num_partitions=NUM_PARTITIONS,
-            **common_kwargs,
-        )
-        for i in range(NUM_PARTITIONS)
+    future_0 = run_ssim_partition.spawn(partition_index=0)
+    future_1 = run_ssim_partition.spawn(partition_index=1)
+    results: list[_PartitionResult | None] = [
+        future_0.get(),
+        future_1.get(),
     ]
-    results: list[_PartitionResult | None] = [f.get() for f in futures]
 
     exit_code = _print_combined_results(results)
-    if sync_generated_to_volume:
-        download_src = _build_generated_volume_relative_path(
-            generated_volume_subdir=resolved_subdir,
-            quality_tier=quality_tier,
-        )
-        local_download_dir = _build_local_generated_download_dir(quality_tier)
-        print(
-            "To download raw generated videos locally, run:\n"
-            f"  modal volume get hf-model-weights {download_src} "
-            f"{local_download_dir}"
-        )
-        _print_local_reference_copy_command(quality_tier)
     if exit_code != 0:
         sys.exit(exit_code)
     print("All SSIM tasks passed.")

fastvideo/tests/nightly/test_e2e_dmd_t2v_crush_smol.py

@@ -3,8 +3,8 @@ from pathlib import Path
 from huggingface_hub import snapshot_download
 import subprocess
 import sys
+from fastvideo.tests.ssim.test_inference_similarity import compute_video_ssim_torchvision
 import shutil
-from fastvideo.tests.utils import compute_video_ssim_torchvision
 
 # Import the training pipeline
 sys.path.append(str(Path(__file__).parent.parent.parent.parent.parent))
@@ -181,4 +181,4 @@ def test_e2e_overfit_single_sample():
 
 
 if __name__ == "__main__":
-    test_e2e_overfit_single_sample()
+    test_e2e_overfit_single_sample()

fastvideo/tests/nightly/test_e2e_i2v_overfit_single_sample.py

@@ -4,7 +4,7 @@ from huggingface_hub import snapshot_download
 import shutil
 import subprocess
 import sys
-from fastvideo.tests.utils import compute_video_ssim_torchvision
+from fastvideo.tests.ssim.test_inference_similarity import compute_video_ssim_torchvision
 
 # Import the training pipeline
 sys.path.append(str(Path(__file__).parent.parent.parent.parent.parent))
@@ -170,4 +170,4 @@ def test_e2e_overfit_single_sample():
 
 
 if __name__ == "__main__":
-    test_e2e_overfit_single_sample()
+    test_e2e_overfit_single_sample()

fastvideo/tests/nightly/test_e2e_overfit_single_sample.py

@@ -4,7 +4,7 @@ from huggingface_hub import snapshot_download
 import shutil
 import subprocess
 import sys
-from fastvideo.tests.utils import compute_video_ssim_torchvision
+from fastvideo.tests.ssim.test_inference_similarity import compute_video_ssim_torchvision
 
 # Import the training pipeline
 sys.path.append(str(Path(__file__).parent.parent.parent.parent.parent))
@@ -175,4 +175,4 @@ def test_e2e_overfit_single_sample():
 
 
 if __name__ == "__main__":
-    test_e2e_overfit_single_sample()
+    test_e2e_overfit_single_sample()

fastvideo/tests/ssim/AGENTS.md

@@ -1,55 +0,0 @@
-# SSIM Directory Guidelines
-
-## Scope
-These instructions apply to everything under `fastvideo/tests/ssim/`.
-
-## Purpose
-SSIM tests are GPU-backed end-to-end quality regression checks. They generate
-videos and compare them against device-specific reference videos.
-
-## Key Files
-- `test_*.py`: SSIM regression tests by model/task.
-- `conftest.py`: optional filtering via `FASTVIDEO_SSIM_MODEL_ID`.
-- `reference_videos/<quality-tier>/<GPU>_reference_videos/`: local cache of
-  references split by quality tier and GPU type.
-- `generated_videos/<quality-tier>/<GPU>_reference_videos/`: local outputs and
-  `*_ssim.json` artifacts (git-ignored).
-- `reference_videos_cli.py`: copy/download/upload/ensure reference videos
-  (including HF sync).
-
-## Run Commands
-- Full SSIM suite: `pytest fastvideo/tests/ssim/ -vs`
-- Full SSIM with full-quality params:
-  `pytest fastvideo/tests/ssim/ -vs --ssim-full-quality`
-- Single test file:
-  `pytest fastvideo/tests/ssim/test_wan_t2v_similarity.py -vs`
-- Single model split:
-  `FASTVIDEO_SSIM_MODEL_ID=<model_id> pytest fastvideo/tests/ssim/test_wan_t2v_similarity.py -vs`
-- Modal orchestrator (CI-style scheduling):
-  `modal run fastvideo/tests/modal/ssim_test.py`
-
-## Authoring Rules
-- Name files `test_<feature>_similarity.py`.
-- Set `REQUIRED_GPUS = <int>` near the top of each test module.
-- For multi-model suites, keep configs in `*_MODEL_TO_PARAMS` dictionaries.
-  The Modal scheduler auto-discovers these keys and runs one subprocess per
-  model id.
-- Keep runs deterministic when possible (fixed prompts/seeds/frames/backend).
-- Persist metrics with `write_ssim_results(...)`.
-- Use `pytest.skip(...)` for unsupported hardware, missing assets, or
-  insufficient GPU count.
-
-## Updating Reference Videos
-1. Run the target SSIM test and inspect generated outputs + scores.
-2. Update references only for intentional behavior changes.
-3. Copy local generated outputs into the target reference folder:
-   `python fastvideo/tests/ssim/reference_videos_cli.py copy-local ...`
-4. Upload updated reference folders to HF:
-   `python fastvideo/tests/ssim/reference_videos_cli.py upload ...`
-5. In the PR, explain why references changed and which GPU type generated
-   them.
-
-## PR Expectations
-- Include exact test commands and pass/fail evidence.
-- If thresholds change, include before/after SSIM numbers and rationale.
-- If references change, call out the source commit/model/backend.

fastvideo/tests/ssim/L40S_reference_videos/TurboWan2.2-I2V-A14B-Diffusers/SLA_ATTN/steps4_An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space reali_ssim.json

@@ -0,0 +1,11 @@
+{
+  "mean_ssim": 0.9805062346988254,
+  "min_ssim": 0.9636241793632507,
+  "max_ssim": 0.9951534271240234,
+  "reference_video": "/mnt/fast-disks/hao_lab/loay/FastVideo/fastvideo/tests/ssim/L40S_reference_videos/TurboWan2.2-I2V-A14B-Diffusers/SLA_ATTN/An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space reali.mp4",
+  "generated_video": "/mnt/fast-disks/hao_lab/loay/FastVideo/fastvideo/tests/ssim/generated_videos/TurboWan2.2-I2V-A14B-Diffusers/SLA_ATTN/An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space reali.mp4",
+  "parameters": {
+    "num_inference_steps": 4,
+    "prompt": "An astronaut hatching from an egg, on the surface of the moon, the darkness and depth of space realised in the background. High quality, ultrarealistic detail and breath-taking movie-like camera shot."
+  }
+}