[feat] Migrate Hunyuan SamplingParam subclasses to profile-based defaults

Replace HunyuanSamplingParam and FastHunyuanSamplingParam subclasses with profile-based defaults. This adds the profile infrastructure (ConfigInfo.default_profile, SamplingParam._from_profile) and the first profile definitions for Hunyuan models, enabling future model families to migrate from subclass overrides to declarative profiles. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
[feat] [3/n] Improve API: extend support to cli (#1226 )
2026-04-15 16:33:37 -07:00 · 2026-04-14 15:20:47 -07:00 · 2026-04-06 10:33:54 -07:00 · 2026-04-06 02:20:26 +00:00 · 2026-04-06 02:03:46 +00:00 · 2026-04-05 18:13:19 -07:00
556 changed files with 38550 additions and 14714 deletions
@@ -2,20 +2,218 @@ 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"
+  - github_commit_status:
+      context: "direct-test-completed"
+    if: build.env("TEST_SCOPE") == "direct"
+
 steps:
-  - label: "pre-commit"
-    command: ".buildkite/scripts/pre_commit.sh"
-    agents:
-      queue: "default"
+  # ============================================================
+  # Direct test: triggered by /test <name> slash command.
+  # Labels match fastcheck/full-suite counterparts so the GitHub
+  # check status overwrites the original failed check.
+  # Only ONE step executes per build (gated by TEST_TYPE).
+  # ============================================================

-  - wait
+  # --- Fastcheck-scope direct tests ---
+    - label: ":microscope: Encoder Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "encoder"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":microscope: VAE Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "vae"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":microscope: Transformer Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "transformer"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":microscope: Kernel Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "kernel_tests"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":microscope: Unit Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "unit_test"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"

-  - 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: 
+  # --- Full-suite-scope direct tests ---
+    - label: ":bar_chart: SSIM Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "ssim"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+          - exit_status: 1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: LoRA Inference Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "inference_lora"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: Training Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "training"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: Distillation DMD Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "distillation_dmd"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: Self-Forcing Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "self_forcing"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: LoRA Training Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "training_lora"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+          - exit_status: 1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: Training Tests VSA"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "training_vsa"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+          - exit_status: 1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: Inference Tests VMoBA"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "inference_vmoba"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: Performance Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "performance"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      agents:
+        queue: "default"
+    - label: ":test_tube: API Server Tests"
+      if: build.env("TEST_SCOPE") == "direct" && build.env("TEST_TYPE") == "api_server"
+      command: "timeout 90m .buildkite/scripts/pr_test.sh"
+      retry:
+        automatic:
+          - exit_status: 128
+            limit: 3
+          - exit_status: -1
+            limit: 2
+      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:
                - "fastvideo/models/encoders/**"
                - "fastvideo/models/loader/**"
                - "fastvideo/tests/encoders/**"
@@ -23,7 +221,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 20m .buildkite/scripts/pr_test.sh"
-                label: "Encoder Tests"
+                label: ":microscope: Encoder Tests"
                env:
                  - TEST_TYPE=encoder
                agents:
@@ -36,7 +234,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 20m .buildkite/scripts/pr_test.sh"
-                label: "VAE Tests"
+                label: ":microscope: VAE Tests"
                env:
                  - TEST_TYPE=vae
                agents:
@@ -51,20 +249,68 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Transformer Tests"
+                label: ":microscope: Transformer Tests"
                env:
                  - TEST_TYPE=transformer
                agents:
                  queue: "default"
-            - path: 
+            - 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:
                - "fastvideo/**/*.py"
                - "pyproject.toml"
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 90m .buildkite/scripts/pr_test.sh"
-                label: "SSIM Tests"
+                label: ":bar_chart: SSIM Tests"
                env:
                  - TEST_TYPE=ssim
+                retry:
+                  automatic:
+                    - exit_status: 1
+                      limit: 2
                agents:
                  queue: "default"
            - path:
@@ -77,7 +323,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 20m .buildkite/scripts/pr_test.sh"
-                label: "LoRA Inference Tests"
+                label: ":test_tube: LoRA Inference Tests"
                env:
                  - TEST_TYPE=inference_lora
                agents:
@@ -88,7 +334,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Training Tests"
+                label: ":test_tube: Training Tests"
                env:
                  - TEST_TYPE=training
                agents:
@@ -99,7 +345,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Distillation DMDTests"
+                label: ":test_tube: Distillation DMD Tests"
                env:
                  - TEST_TYPE=distillation_dmd
                agents:
@@ -111,7 +357,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Self-Forcing Tests"
+                label: ":test_tube: Self-Forcing Tests"
                env:
                  - TEST_TYPE=self_forcing
                agents:
@@ -122,9 +368,13 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "LoRA Training Tests"
+                label: ":test_tube: LoRA Training Tests"
                env:
                  - TEST_TYPE=training_lora
+                retry:
+                  automatic:
+                    - exit_status: 1
+                      limit: 2
                agents:
                  queue: "default"
            - path:
@@ -134,20 +384,15 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Training Tests VSA"
+                label: ":test_tube: Training Tests VSA"
                env:
                  - TEST_TYPE=training_vsa
+                retry:
+                  automatic:
+                    - exit_status: 1
+                      limit: 2
                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"
@@ -155,20 +400,9 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Inference Tests VMoBA"
-                env:  
-                  - TEST_TYPE=inference_vmoba
-                agents:
-                  queue: "default"
-            - path:
-                - "fastvideo/**"
-                - "pyproject.toml"
-                - "docker/Dockerfile.python3.12"
-              config:
-                command: "timeout 15m .buildkite/scripts/pr_test.sh"
-                label: "Unit Tests"
+                label: ":test_tube: Inference Tests VMoBA"
                env:
-                  - TEST_TYPE=unit_test
+                  - TEST_TYPE=inference_vmoba
                agents:
                  queue: "default"
            - path:
@@ -184,7 +418,7 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 30m .buildkite/scripts/pr_test.sh"
-                label: "Performance Tests"
+                label: ":test_tube: Performance Tests"
                env:
                  - TEST_TYPE=performance
                agents:
@@ -197,19 +431,8 @@ steps:
                - "docker/Dockerfile.python3.12"
              config:
                command: "timeout 30m .buildkite/scripts/pr_test.sh"
-                label: "API Server Tests"
+                label: ":test_tube: 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"
@@ -59,7 +59,11 @@ if [ -z "${TEST_TYPE:-}" ]; then
 fi
 log "Test type: $TEST_TYPE"

-MODAL_ENV="BUILDKITE_REPO=$BUILDKITE_REPO BUILDKITE_COMMIT=$BUILDKITE_COMMIT BUILDKITE_PULL_REQUEST=$BUILDKITE_PULL_REQUEST IMAGE_VERSION=$IMAGE_VERSION"
+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"

 case "$TEST_TYPE" in
    "encoder")
@@ -1,3 +1,22 @@
+<!--
+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. -->
@@ -0,0 +1,316 @@
+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)\\]"
+      - "#approved-reviews-by>=1"
+      - check-success~=pre-commit
+      - check-success=fastcheck-passed
+      - check-success=full-suite-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-update when ready
+    conditions:
+      - label=ready
+      - "#approved-reviews-by>=1"
+      - -conflict
+      - -closed
+      - -draft
+    actions:
+      update: {}
+
+  # ============================================================
+  # 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.
+
+merge_protections_settings:
+  reporting_method: check-runs
@@ -1,249 +0,0 @@
-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()
@@ -1,90 +0,0 @@
-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()
@@ -0,0 +1,80 @@
+name: Aggregate Test Status
+
+on:
+  status:
+
+permissions:
+  statuses: write
+
+jobs:
+  aggregate:
+    if: >-
+      github.event.context == 'direct-test-completed'
+      && github.event.state == 'success'
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check and update aggregate status
+        uses: actions/github-script@60a0d83039c74a4aee543508d2ffcb1c3799cdea # v7.0.1
+        with:
+          script: |
+            const sha = context.payload.sha;
+
+            const { data } = await github.rest.repos.getCombinedStatusForRef({
+              owner: context.repo.owner,
+              repo: context.repo.repo,
+              ref: sha,
+              per_page: 100,
+            });
+
+            const bkStatuses = data.statuses.filter(
+              s => s.context.startsWith('buildkite/ci/')
+            );
+
+            const FASTCHECK_PREFIX = 'buildkite/ci/microscope-';
+            const FULL_SUITE_PREFIXES = [
+              'buildkite/ci/test-tube-',
+              'buildkite/ci/bar-chart-',
+            ];
+
+            const fastcheck = bkStatuses.filter(
+              s => s.context.startsWith(FASTCHECK_PREFIX)
+            );
+            const fullSuite = bkStatuses.filter(
+              s => FULL_SUITE_PREFIXES.some(p => s.context.startsWith(p))
+            );
+
+            if (
+              fastcheck.length > 0
+              && fastcheck.every(s => s.state === 'success')
+            ) {
+              core.info(
+                `All ${fastcheck.length} fastcheck tests passed — updating fastcheck-passed`
+              );
+              await github.rest.repos.createCommitStatus({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                sha,
+                state: 'success',
+                context: 'fastcheck-passed',
+                description:
+                  `All ${fastcheck.length} fastcheck tests passed`,
+              });
+            }
+
+            if (
+              fullSuite.length > 0
+              && fullSuite.every(s => s.state === 'success')
+            ) {
+              core.info(
+                `All ${fullSuite.length} full suite tests passed — updating full-suite-passed`
+              );
+              await github.rest.repos.createCommitStatus({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                sha,
+                state: 'success',
+                context: 'full-suite-passed',
+                description:
+                  `All ${fullSuite.length} full suite tests passed`,
+              });
+            }
@@ -0,0 +1,32 @@
+name: pre-commit
+
+on:
+  pull_request:
+    branches: [main]
+  workflow_call:
+    inputs:
+      ref:
+        description: 'Git ref to checkout (defaults to github.ref)'
+        required: false
+        type: string
+
+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
+      with:
+        ref: ${{ inputs.ref || '' }}
+    - 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
@@ -0,0 +1,271 @@
+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
+        id: label
+        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;
+            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',
+            });
+            const { data: pr } = await github.rest.pulls.get({ owner, repo, pull_number: prNumber });
+            core.setOutput('pr_sha', pr.head.sha);
+            core.setOutput('pr_branch', pr.head.ref);
+            core.setOutput('pr_number', String(prNumber));
+
+      - name: Trigger Full Suite
+        if: steps.perm.outputs.has_write == 'true'
+        env:
+          BUILDKITE_API_TOKEN: ${{ secrets.BUILDKITE_API_TOKEN }}
+          PR_SHA: ${{ steps.label.outputs.pr_sha }}
+          PR_BRANCH: ${{ steps.label.outputs.pr_branch }}
+          PR_NUMBER: ${{ steps.label.outputs.pr_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 #${PR_NUMBER} (via /merge)" \
+              --argjson pr_id "$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)
+                }
+              }')"
+
+  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);
+
+      - name: React to comment
+        if: steps.perm.outputs.has_write == 'true'
+        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',
+            });
+
+  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
+    with:
+      ref: refs/pull/${{ github.event.issue.number }}/merge
+
+  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: 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: $pr_number
+                }
+              }')"
@@ -0,0 +1,83 @@
+name: Trigger Full Suite
+
+on:
+  pull_request_target:
+    types: [labeled, synchronize]
+
+permissions:
+  contents: read
+  pull-requests: read
+
+concurrency:
+  group: full-suite-${{ github.event.pull_request.number }}
+  cancel-in-progress: false
+
+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(try (.env.TEST_SCOPE == "full") catch false) | .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)
+                }
+              }')"
@@ -0,0 +1,65 @@
+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');
+            }
@@ -0,0 +1,51 @@
+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!
@@ -0,0 +1,56 @@
+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)
@@ -32,7 +32,7 @@ permissions:
 jobs:
  build-python-3-10:
    if: ${{ github.event.inputs.python_3_10 == 'true' }}
-    uses: ./.github/workflows/build-image-template.yml
+    uses: ./.github/workflows/_template-build-image.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/build-image-template.yml
+    uses: ./.github/workflows/_template-build-image.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/build-image-template.yml
+    uses: ./.github/workflows/_template-build-image.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/build-image-template.yml
+    uses: ./.github/workflows/_template-build-image.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
@@ -7,14 +7,14 @@ on:
      - 'docs/**'
      - 'mkdocs.yml'
      - 'requirements-mkdocs.txt'
-      - '.github/workflows/docs.yml'
+      - '.github/workflows/infra-docs.yml'
  pull_request:
    branches: [ main ]
    paths:
      - 'docs/**'
      - 'mkdocs.yml'
      - 'requirements-mkdocs.txt'
-      - '.github/workflows/docs.yml'
+      - '.github/workflows/infra-docs.yml'

 permissions:
  contents: read
@@ -0,0 +1,17 @@
+{
+  "problemMatcher": [
+    {
+      "owner": "ruff",
+      "pattern": [
+        {
+          "regexp": "^(.+):(\\d+):(\\d+): (\\w+) (.+)$",
+          "file": 1,
+          "line": 2,
+          "column": 3,
+          "code": 4,
+          "message": 5
+        }
+      ]
+    }
+  ]
+}
@@ -1,338 +0,0 @@
-name: PR Test
-
-on:
-  push:
-    branches: [main]
-    paths:
-      - "fastvideo/**/*.py"
-      - ".github/workflows/pr-test.yml"
-  pull_request:
-    branches: [main]
-    types: [opened, ready_for_review, synchronize, reopened]
-    paths:
-      - "fastvideo/**/*.py"
-      - ".github/workflows/pr-test.yml"
-      - "pyproject.toml"
-      - "docker/Dockerfile.python3.12"
-      - "csrc/**"
-  workflow_dispatch:
-    inputs:
-      run_encoder_test:
-        description: "Run encoder-test"
-        required: false
-        default: false
-        type: boolean
-      run_vae_test:
-        description: "Run vae-test"
-        required: false
-        default: false
-        type: boolean
-      run_transformer_test:
-        description: "Run transformer-test"
-        required: false
-        default: false
-        type: boolean
-      run_ssim_test:
-        description: "Run ssim-test"
-        required: false
-        default: false
-        type: boolean
-      run_training_test:
-        description: "Run training-test"
-        required: false
-        default: false
-        type: boolean
-      run_training_test_VSA:
-        description: "Run training-test-VSA"
-        required: false
-        default: false
-        type: boolean
-      run_precision_test_VSA:
-        description: "Run precision-test-VSA"
-        required: false
-        default: false
-        type: boolean
-      run_unit_test:
-        description: "Run unit-test"
-        required: false
-        default: false
-        type: boolean
-
-env:
-  PYTHONUNBUFFERED: "1"
-
-
-concurrency:
-  group: pr-test-${{ 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
@@ -1,18 +0,0 @@
-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
@@ -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:
@@ -1,94 +0,0 @@
-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 
@@ -1,31 +0,0 @@
-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
@@ -1,257 +0,0 @@
-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/
@@ -33,6 +33,7 @@ env
 **.txt
 *.log
 weights/
+logs/

 # SSIM test outputs
 fastvideo/tests/ssim/generated_videos/
@@ -84,6 +85,10 @@ docs/distillation/examples/
 dmd_t2v_output/
 preprocess_output_text/

+# Next.js / Node artifacts under ui/: see ui/.gitignore
+
 .claude/
 .codex/
+.sisyphus/
 openspec/
+fastvideo/tests/ssim/reference_videos/**
@@ -0,0 +1 @@
+WRN 2026-03-26T13:46:33.469 ?.19646    server_start:193: Failed to start server: operation not permitted: /var/folders/z_/h_6myyk14d1b7z87z3vy4mjh0000gn/T/nvim.dsynkd/iSe0el/nvim.19646.0
@@ -18,10 +18,8 @@ exclude: |
    fastvideo/train\.py|
    fastvideo/utils/.*|
    examples/.*|
-    .github/workflows/fastvideo-publish.yml|
-    .github/workflows/sta-publish.yml|
-    .github/workflows/vsa-publish.yml|
-    .github/workflows/build-image-template.yml|
+    .github/workflows/publish-fastvideo.yml|
+    .github/workflows/_template-build-image.yml|
    docs/source/inference/support_matrix.md
  )
 repos:
@@ -0,0 +1 @@
+3.12
@@ -6,16 +6,17 @@
     | <a href="https://hao-ai-lab.github.io/FastVideo"><b>Documentation</b></a> | <a href="https://hao-ai-lab.github.io/FastVideo/inference/inference_quick_start/"><b> Quick Start</b></a> | <a href="https://github.com/hao-ai-lab/FastVideo/discussions/982"  target="_blank"><b>Weekly Dev Meeting</b></a>  | 🟣💬 <a href="https://join.slack.com/t/fastvideo/shared_invite/zt-3f4lao1uq-u~Ipx6Lt4J27AlD2y~IdLQ" target="_blank"> <b>Slack</b> </a> |  🟣💬 <a href="https://github.com/hao-ai-lab/FastVideo/discussions/1097" target="_blank"> <b> WeChat </b> </a> |
 </p>

-**FastVideo is a unified post-training and inference framework for accelerated video generation.**
+**FastVideo is a unified post-training and real-time inference framework for accelerated video generation.**

 ## NEWS
-
- `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)
+- `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/).

 ### More News

- `2025/06/14`: Release finetuning and inference code for [VSA](https://arxiv.org/pdf/2505.13389)
+- `2025/06/14`: Release finetuning and inference code for [VSA](https://arxiv.org/pdf/2505.13389).
 - `2025/04/24`: [FastVideo V1](https://hao-ai-lab.github.io/blogs/fastvideo/) is released!
 - `2025/02/18`: Release the inference code for [Sliding Tile Attention](https://hao-ai-lab.github.io/blogs/sta/).

@@ -1,15 +1,10 @@
 try:
-    from .comfyui.video_generator.nodes import (NODE_CLASS_MAPPINGS,
-                                                NODE_DISPLAY_NAME_MAPPINGS)
+    from .comfyui.video_generator.nodes import (NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS)
    WEB_DIRECTORY = "./web"
-    __all__ = [
-        'NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY'
-    ]
+    __all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY']
 except ImportError:
    # ComfyUI environment not available, skip comfyui imports
    NODE_CLASS_MAPPINGS = {}
    NODE_DISPLAY_NAME_MAPPINGS = {}
    WEB_DIRECTORY = "./web"
-    __all__ = [
-        'NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY'
-    ]
+    __all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY']
@@ -13,8 +13,7 @@ from .fvd import (
    compute_statistics,
    FVDConfig,
 )
-from .feature_extractors import (BaseFeatureExtractor, I3DFeatureExtractor,
-                                 load_extractor)
+from .feature_extractors import (BaseFeatureExtractor, I3DFeatureExtractor, load_extractor)
 from .video_utils import (
    load_video_auto,
    sample_clips_from_video,
@@ -5,18 +5,11 @@ from .fvd import compute_fvd_with_config, FVDConfig


 def main() -> int:
-    parser = argparse.ArgumentParser(
-        description='Compute Fréchet Video Distance (FVD)')
+    parser = argparse.ArgumentParser(description='Compute Fréchet Video Distance (FVD)')

    # Required arguments
-    parser.add_argument('--real-path',
-                        type=str,
-                        required=True,
-                        help='Path to real videos')
-    parser.add_argument('--gen-path',
-                        type=str,
-                        required=True,
-                        help='Path to generated videos')
+    parser.add_argument('--real-path', type=str, required=True, help='Path to real videos')
+    parser.add_argument('--gen-path', type=str, required=True, help='Path to generated videos')

    # Extractor selection
    parser.add_argument('--extractor',
@@ -26,42 +19,19 @@ def main() -> int:
                        help='Feature extractor model to use (default: i3d)')

    # Standard args
-    parser.add_argument('--seed',
-                        type=int,
-                        default=None,
-                        help='Random seed for reproducibility')
+    parser.add_argument('--seed', type=int, default=None, help='Random seed for reproducibility')
    parser.add_argument('--protocol',
                        type=str,
                        default=None,
                        choices=['fvd2048_16f', 'fvd2048_128f', 'quick_test'],
                        help='Use standard protocol (overrides other settings)')
-    parser.add_argument('--num-videos',
-                        type=int,
-                        default=2048,
-                        help='Number of videos to use')
-    parser.add_argument('--num-frames',
-                        type=int,
-                        default=16,
-                        help='Number of frames per clip')
-    parser.add_argument('--clip-strategy',
-                        type=str,
-                        default='beginning',
-                        help='Clip sampling strategy')
-    parser.add_argument('--batch-size',
-                        type=int,
-                        default=32,
-                        help='Batch size for feature extraction')
-    parser.add_argument('--device',
-                        type=str,
-                        default='cuda',
-                        help='Device to use (cuda or cpu)')
-    parser.add_argument('--cache-real-features',
-                        type=str,
-                        default=None,
-                        help='Path to cache real video features')
-    parser.add_argument('--quiet',
-                        action='store_true',
-                        help='Suppress progress output')
+    parser.add_argument('--num-videos', type=int, default=2048, help='Number of videos to use')
+    parser.add_argument('--num-frames', type=int, default=16, help='Number of frames per clip')
+    parser.add_argument('--clip-strategy', type=str, default='beginning', help='Clip sampling strategy')
+    parser.add_argument('--batch-size', type=int, default=32, help='Batch size for feature extraction')
+    parser.add_argument('--device', type=str, default='cuda', help='Device to use (cuda or cpu)')
+    parser.add_argument('--cache-real-features', type=str, default=None, help='Path to cache real video features')
+    parser.add_argument('--quiet', action='store_true', help='Suppress progress output')

    args = parser.parse_args()

@@ -22,8 +22,7 @@ class BaseFeatureExtractor(ABC, nn.Module):

    def __init__(self, device: str = 'cuda'):
        super().__init__()
-        self.device = torch.device(
-            device if torch.cuda.is_available() else 'cpu')
+        self.device = torch.device(device if torch.cuda.is_available() else 'cpu')

    @property
    @abstractmethod
@@ -53,10 +52,7 @@ class BaseFeatureExtractor(ABC, nn.Module):
        pass

    @torch.no_grad()
-    def extract_features(self,
-                         videos: torch.Tensor,
-                         batch_size: int = 32,
-                         verbose: bool = True) -> torch.Tensor:
+    def extract_features(self, videos: torch.Tensor, batch_size: int = 32, verbose: bool = True) -> torch.Tensor:
        """
        Extract features for a large tensor of videos by batching.
        """
@@ -65,9 +61,7 @@ class BaseFeatureExtractor(ABC, nn.Module):

        iterator = range(0, N, batch_size)
        if verbose:
-            iterator = tqdm(
-                iterator,
-                desc=f"Extracting features ({self.__class__.__name__})")
+            iterator = tqdm(iterator, desc=f"Extracting features ({self.__class__.__name__})")

        for i in iterator:
            batch = videos[i:i + batch_size].to(self.device)
@@ -95,9 +89,7 @@ class I3DFeatureExtractor(BaseFeatureExtractor):

    def _load_model(self) -> torch.nn.Module:
        try:
-            model_path = hf_hub_download(repo_id=self.REPO_ID,
-                                         filename=self.MODEL_FILENAME,
-                                         cache_dir=self.cache_dir)
+            model_path = hf_hub_download(repo_id=self.REPO_ID, filename=self.MODEL_FILENAME, cache_dir=self.cache_dir)
            return torch.jit.load(model_path, map_location=self.device)
        except Exception as e:
            raise RuntimeError(f"Failed to load I3D model: {e}") from e
@@ -119,10 +111,7 @@ class I3DFeatureExtractor(BaseFeatureExtractor):
        # Resize to 224x224
        if H != 224 or W != 224:
            videos = videos.reshape(B * T, C, H, W)
-            videos = F.interpolate(videos,
-                                   size=(224, 224),
-                                   mode='bilinear',
-                                   align_corners=False)
+            videos = F.interpolate(videos, size=(224, 224), mode='bilinear', align_corners=False)
            videos = videos.reshape(B, T, C, 224, 224)

        # [B, T, C, H, W] -> [B, C, T, H, W]
@@ -131,21 +120,15 @@ class I3DFeatureExtractor(BaseFeatureExtractor):
    def extract_features_batch(self, videos: torch.Tensor) -> torch.Tensor:
        batch = self.preprocess(videos)
        # TorchScript I3D returns raw logits when return_features=True
-        return self.model(batch,
-                          rescale=False,
-                          resize=False,
-                          return_features=True)
+        return self.model(batch, rescale=False, resize=False, return_features=True)


 # 2. CLIP Extractor (Semantic/Content Quality)
 class CLIPFeatureExtractor(BaseFeatureExtractor):

-    def __init__(self,
-                 device: str = 'cuda',
-                 model_name: str = "openai/clip-vit-base-patch32"):
+    def __init__(self, device: str = 'cuda', model_name: str = "openai/clip-vit-base-patch32"):
        if not TRANSFORMERS_AVAILABLE:
-            raise ImportError(
-                "Please install transformers: pip install transformers")
+            raise ImportError("Please install transformers: pip install transformers")
        super().__init__(device)
        self.processor = CLIPProcessor.from_pretrained(model_name)
        self.model = CLIPModel.from_pretrained(model_name).to(self.device)
@@ -172,9 +155,7 @@ class CLIPFeatureExtractor(BaseFeatureExtractor):
        images = videos.view(B * T, C, H, W)

        # HF Processor
-        inputs = self.processor(images=images,
-                                return_tensors="pt",
-                                padding=True)
+        inputs = self.processor(images=images, return_tensors="pt", padding=True)
        inputs = {k: v.to(self.device) for k, v in inputs.items()}

        # Extract features [B*T, Dim]
@@ -188,24 +169,15 @@ class CLIPFeatureExtractor(BaseFeatureExtractor):
 # 3. VideoMAE Extractor (Structure/Motion Quality)
 class VideoMAEFeatureExtractor(BaseFeatureExtractor):

-    def __init__(self,
-                 device: str = 'cuda',
-                 model_name: str = "MCG-NJU/videomae-base"):
+    def __init__(self, device: str = 'cuda', model_name: str = "MCG-NJU/videomae-base"):
        if not TRANSFORMERS_AVAILABLE:
-            raise ImportError(
-                "Please install transformers: pip install transformers")
+            raise ImportError("Please install transformers: pip install transformers")
        super().__init__(device)
        self.model = VideoMAEModel.from_pretrained(model_name).to(self.device)
        self.model.eval()

-        self.register_buffer(
-            'mean',
-            torch.tensor([0.485, 0.456, 0.406],
-                         device=self.device).view(1, 1, 3, 1, 1))
-        self.register_buffer(
-            'std',
-            torch.tensor([0.229, 0.224, 0.225],
-                         device=self.device).view(1, 1, 3, 1, 1))
+        self.register_buffer('mean', torch.tensor([0.485, 0.456, 0.406], device=self.device).view(1, 1, 3, 1, 1))
+        self.register_buffer('std', torch.tensor([0.229, 0.224, 0.225], device=self.device).view(1, 1, 3, 1, 1))

    @property
    def feature_dim(self) -> int:
@@ -221,10 +193,7 @@ class VideoMAEFeatureExtractor(BaseFeatureExtractor):
        # 1. Resize to 224x224
        if H != 224 or W != 224:
            videos = videos.view(B * T, C, H, W)
-            videos = F.interpolate(videos,
-                                   size=(224, 224),
-                                   mode='bilinear',
-                                   align_corners=False)
+            videos = F.interpolate(videos, size=(224, 224), mode='bilinear', align_corners=False)
            videos = videos.view(B, T, C, 224, 224)

        # 2. Normalize to [0, 1]
@@ -260,5 +229,4 @@ def load_extractor(name: str, device: str = 'cuda') -> BaseFeatureExtractor:
    elif name == 'videomae':
        return VideoMAEFeatureExtractor(device)
    else:
-        raise ValueError(
-            f"Unknown extractor: {name}. Options: i3d, clip, videomae")
+        raise ValueError(f"Unknown extractor: {name}. Options: i3d, clip, videomae")
@@ -36,8 +36,7 @@ def compute_frechet_distance(mu1: np.ndarray,

    if np.iscomplexobj(covmean):
        if not np.allclose(np.diagonal(covmean).imag, 0, atol=1e-3):
-            print(
-                f"Warning: Imaginary component: {np.max(np.abs(covmean.imag))}")
+            print(f"Warning: Imaginary component: {np.max(np.abs(covmean.imag))}")
        covmean = covmean.real

    trace_product = np.trace(covmean)
@@ -67,8 +66,7 @@ class FVDConfig:
    temporal_stride: int = 1  # For sliding window clips

    # Data processing
-    video_extensions: list[str] = field(
-        default_factory=lambda: ['.mp4', '.avi', '.mov', '.mkv'])
+    video_extensions: list[str] = field(default_factory=lambda: ['.mp4', '.avi', '.mov', '.mkv'])
    support_frame_dirs: bool = True

    # Computation
@@ -88,25 +86,17 @@ class FVDConfig:
    @classmethod
    def fvd2048_16f(cls) -> 'FVDConfig':
        """Standard FVD protocol: 2048 videos, 16 frames, beginning clip."""
-        return cls(num_videos=2048,
-                   num_frames_per_clip=16,
-                   clip_strategy='beginning',
-                   use_streaming=True)
+        return cls(num_videos=2048, num_frames_per_clip=16, clip_strategy='beginning', use_streaming=True)

    @classmethod
    def fvd2048_128f(cls) -> 'FVDConfig':
        """Long video protocol: 2048 videos, 128 frames."""
-        return cls(num_videos=2048,
-                   num_frames_per_clip=128,
-                   clip_strategy='beginning',
-                   use_streaming=True)
+        return cls(num_videos=2048, num_frames_per_clip=128, clip_strategy='beginning', use_streaming=True)

    @classmethod
    def quick_test(cls) -> 'FVDConfig':
        """Quick test config: 100 videos, 16 frames."""
-        return cls(num_videos=100,
-                   num_frames_per_clip=16,
-                   clip_strategy='beginning')
+        return cls(num_videos=100, num_frames_per_clip=16, clip_strategy='beginning')

    def to_dict(self) -> dict:
        """Export config to dict for logging"""
@@ -196,14 +186,10 @@ def load_or_compute_features(videos: str | Path | torch.Tensor,
            max_features = config.num_videos * config.num_clips_per_video

            if len(features) < max_features:
-                print(
-                    f"WARNING: Cache has {len(features)} features but need {max_features}"
-                )
+                print(f"WARNING: Cache has {len(features)} features but need {max_features}")
                print("Cached features insufficient - will recompute...")
            elif len(features) > max_features:
-                print(
-                    f"Using {max_features} features from cache (truncated from {len(features)})"
-                )
+                print(f"Using {max_features} features from cache (truncated from {len(features)})")
                features = features[:max_features]
                return features
            else:
@@ -215,17 +201,16 @@ def load_or_compute_features(videos: str | Path | torch.Tensor,
    if isinstance(videos, (str | Path)):
        target_size = (224, 224) if config.resize_before_extraction else None

-        video_generator = load_video_clips_streaming(
-            videos,
-            num_frames=config.num_frames_per_clip,
-            max_videos=config.num_videos,
-            clip_strategy=config.clip_strategy,
-            frame_stride=config.frame_stride,
-            num_clips_per_video=config.num_clips_per_video,
-            video_extensions=config.video_extensions,
-            support_frame_dirs=config.support_frame_dirs,
-            target_size=target_size,
-            verbose=True)
+        video_generator = load_video_clips_streaming(videos,
+                                                     num_frames=config.num_frames_per_clip,
+                                                     max_videos=config.num_videos,
+                                                     clip_strategy=config.clip_strategy,
+                                                     frame_stride=config.frame_stride,
+                                                     num_clips_per_video=config.num_clips_per_video,
+                                                     video_extensions=config.video_extensions,
+                                                     support_frame_dirs=config.support_frame_dirs,
+                                                     target_size=target_size,
+                                                     verbose=True)

        max_clips = config.num_videos * config.num_clips_per_video
        features = extract_features_streaming(video_generator,
@@ -235,17 +220,14 @@ def load_or_compute_features(videos: str | Path | torch.Tensor,
                                              verbose=True)
    else:
        print(f"Extracting features from {len(videos)} video tensors...")
-        features = extractor.extract_features(videos,
-                                              batch_size=config.batch_size,
-                                              verbose=True)
+        features = extractor.extract_features(videos, batch_size=config.batch_size, verbose=True)
        features = features.numpy()

    # Validate feature count
    expected_count = config.num_videos * config.num_clips_per_video
    if len(features) < expected_count:
-        raise ValueError(
-            f"ERROR: Only extracted {len(features)} features, but need {expected_count}!\n"
-            f"Found fewer videos than expected. Check your video directory.")
+        raise ValueError(f"ERROR: Only extracted {len(features)} features, but need {expected_count}!\n"
+                         f"Found fewer videos than expected. Check your video directory.")
    elif len(features) > expected_count:
        print(f"Truncating {len(features)} features to {expected_count}")
        features = features[:expected_count]
@@ -312,9 +294,7 @@ def compute_fvd_with_config(real_videos: str | Path | torch.Tensor,

    # Initialize Extractor using Factory
    if verbose:
-        print(
-            f"\nInitializing {config.extractor_model.upper()} model on {config.device}..."
-        )
+        print(f"\nInitializing {config.extractor_model.upper()} model on {config.device}...")

    extractor = load_extractor(config.extractor_model, device=config.device)

@@ -324,12 +304,11 @@ def compute_fvd_with_config(real_videos: str | Path | torch.Tensor,
        print("Extracting REAL video features...")
        print(f"{'='*70}")

-    real_features = load_or_compute_features(
-        videos=real_videos,
-        extractor=extractor,
-        config=config,
-        cache_path=config.cache_real_features,
-        cache_name="real_features")
+    real_features = load_or_compute_features(videos=real_videos,
+                                             extractor=extractor,
+                                             config=config,
+                                             cache_path=config.cache_real_features,
+                                             cache_name="real_features")

    if verbose:
        print(f"\n{'='*70}")
@@ -19,16 +19,12 @@ class I3DFeatureExtractor(nn.Module):
    REPO_ID = 'flateon/FVD-I3D-torchscript'
    MODEL_FILENAME = 'i3d_torchscript.pt'

-    def __init__(self,
-                 device: str = 'cuda',
-                 cache_dir: str | Path | None = None):
+    def __init__(self, device: str = 'cuda', cache_dir: str | Path | None = None):
        super().__init__()

        self.device_str = device
        if device == 'cuda' and not torch.cuda.is_available():
-            print(
-                "Warning: CUDA requested but not available – falling back to CPU"
-            )
+            print("Warning: CUDA requested but not available – falling back to CPU")
            self.device = torch.device('cpu')
        else:
            self.device = torch.device(device)
@@ -51,9 +47,7 @@ class I3DFeatureExtractor(nn.Module):

        try:
            # Download model from Hugging Face Hub
-            model_path = hf_hub_download(repo_id=self.REPO_ID,
-                                         filename=self.MODEL_FILENAME,
-                                         cache_dir=self.cache_dir)
+            model_path = hf_hub_download(repo_id=self.REPO_ID, filename=self.MODEL_FILENAME, cache_dir=self.cache_dir)

            # Load directly to chosen device
            model = torch.jit.load(model_path, map_location=self.device)
@@ -61,10 +55,9 @@ class I3DFeatureExtractor(nn.Module):
            return model

        except Exception as e:
-            raise RuntimeError(
-                f"Failed to load I3D model from Hugging Face Hub. Error: {e}\n"
-                f"Ensure you have internet connection and huggingface_hub installed:\n"
-                f"pip install huggingface_hub") from e
+            raise RuntimeError(f"Failed to load I3D model from Hugging Face Hub. Error: {e}\n"
+                               f"Ensure you have internet connection and huggingface_hub installed:\n"
+                               f"pip install huggingface_hub") from e

    def preprocess(self, videos: torch.Tensor) -> torch.Tensor:
        """
@@ -88,10 +81,7 @@ class I3DFeatureExtractor(nn.Module):
        # Resize to 224x224 if needed
        if H != 224 or W != 224:
            videos = videos.reshape(B * T, C, H, W)
-            videos = F.interpolate(videos,
-                                   size=(224, 224),
-                                   mode='bilinear',
-                                   align_corners=False)
+            videos = F.interpolate(videos, size=(224, 224), mode='bilinear', align_corners=False)
            videos = videos.reshape(B, T, C, 224, 224)

        # Convert to [B, C, T, H, W] format
@@ -100,10 +90,7 @@ class I3DFeatureExtractor(nn.Module):
        return videos

    @torch.no_grad()
-    def extract_features(self,
-                         videos: torch.Tensor,
-                         batch_size: int = 32,
-                         verbose: bool = True) -> torch.Tensor:
+    def extract_features(self, videos: torch.Tensor, batch_size: int = 32, verbose: bool = True) -> torch.Tensor:
        """
        Extract I3D features
        
@@ -127,16 +114,11 @@ class I3DFeatureExtractor(nn.Module):
            batch = self.preprocess(batch)  # Now returns [B, C, T, H, W]

            # Use the HF model without rescale/resize (we handle it in preprocess)
-            features = self.model(batch,
-                                  rescale=False,
-                                  resize=False,
-                                  return_features=True)
+            features = self.model(batch, rescale=False, resize=False, return_features=True)

            all_features.append(features.cpu())

        return torch.cat(all_features, dim=0)

-    def __call__(self,
-                 videos: torch.Tensor,
-                 batch_size: int = 32) -> torch.Tensor:
+    def __call__(self, videos: torch.Tensor, batch_size: int = 32) -> torch.Tensor:
        return self.extract_features(videos, batch_size=batch_size)
@@ -36,10 +36,7 @@ def main() -> None:
                cache_real_features=str(script_dir / f'fvd-cache/{model_name}'),
            )

-            results = compute_fvd_with_config(real_dir,
-                                              gen_dir,
-                                              cfg,
-                                              verbose=False)
+            results = compute_fvd_with_config(real_dir, gen_dir, cfg, verbose=False)
            print(f"FVD: {results['fvd']}\nModel: {results['model']}")

        except Exception as e:
@@ -59,10 +59,7 @@ def validate_fvd(subset_a: Path, subset_b: Path, num_videos: int):
    print("TEST 1: Identity Test")
    print("=" * 70)

-    result1 = compute_fvd_with_config(real_videos=str(subset_a),
-                                      gen_videos=str(subset_a),
-                                      config=config,
-                                      verbose=False)
+    result1 = compute_fvd_with_config(real_videos=str(subset_a), gen_videos=str(subset_a), config=config, verbose=False)
    fvd_identity = result1['fvd']
    print(f"\nIdentity FVD: {fvd_identity:.2f}")

@@ -70,10 +67,7 @@ def validate_fvd(subset_a: Path, subset_b: Path, num_videos: int):
    print("TEST 2: Real vs Real")
    print("=" * 70)

-    result2 = compute_fvd_with_config(real_videos=str(subset_a),
-                                      gen_videos=str(subset_b),
-                                      config=config,
-                                      verbose=False)
+    result2 = compute_fvd_with_config(real_videos=str(subset_a), gen_videos=str(subset_b), config=config, verbose=False)
    fvd_real = result2['fvd']
    print(f"\nReal vs Real FVD: {fvd_real:.2f}")

@@ -87,9 +81,7 @@ def validate_fvd(subset_a: Path, subset_b: Path, num_videos: int):
 def main() -> None:
    bair_dir = Path('benchmarks/data/bair_full_videos')

-    subset_a, subset_b, count = split_videos(bair_dir,
-                                             n_per_subset=128,
-                                             seed=42)
+    subset_a, subset_b, count = split_videos(bair_dir, n_per_subset=128, seed=42)
    validate_fvd(subset_a, subset_b, count)


@@ -54,8 +54,7 @@ def _load_video_cv2(video_path: str | Path,
            raise RuntimeError(f"Video has 0 frames: {video_path}")

        frames = np.stack(frames)  # [T, H, W, C]
-        frames = torch.from_numpy(frames).permute(0, 3, 1,
-                                                  2).float()  # [T, C, H, W]
+        frames = torch.from_numpy(frames).permute(0, 3, 1, 2).float()  # [T, C, H, W]
        return frames

    if total_frames == 0:
@@ -63,14 +62,11 @@ def _load_video_cv2(video_path: str | Path,

    # Determine frame indices for sampling
    if total_frames < num_frames:
-        frame_indices = list(range(
-            total_frames)) + [total_frames - 1] * (num_frames - total_frames)
+        frame_indices = list(range(total_frames)) + [total_frames - 1] * (num_frames - total_frames)
    elif sample_strategy == 'uniform':
-        frame_indices = np.linspace(0, total_frames - 1, num_frames,
-                                    dtype=int).tolist()
+        frame_indices = np.linspace(0, total_frames - 1, num_frames, dtype=int).tolist()
    elif sample_strategy == 'random':
-        frame_indices = sorted(
-            np.random.choice(total_frames, num_frames, replace=False))
+        frame_indices = sorted(np.random.choice(total_frames, num_frames, replace=False))
    else:
        raise ValueError(f"Unknown sample_strategy: {sample_strategy}")

@@ -94,17 +90,15 @@ def _load_video_cv2(video_path: str | Path,
    cap.release()

    frames = np.stack(frames)  # [T, H, W, C]
-    frames = torch.from_numpy(frames).permute(0, 3, 1,
-                                              2).float()  # [T, C, H, W]
+    frames = torch.from_numpy(frames).permute(0, 3, 1, 2).float()  # [T, C, H, W]

    return frames


-def _load_video_from_frames(
-        frame_dir: str | Path,
-        num_frames: int | None = 16,
-        sample_strategy: str = 'uniform',
-        frame_extensions: list[str] | None = None) -> torch.Tensor:
+def _load_video_from_frames(frame_dir: str | Path,
+                            num_frames: int | None = 16,
+                            sample_strategy: str = 'uniform',
+                            frame_extensions: list[str] | None = None) -> torch.Tensor:
    """
    Load video from directory of frame images.
    
@@ -131,9 +125,7 @@ def _load_video_from_frames(
        frame_files.extend(frame_dir.glob(f"*{ext}"))

    if len(frame_files) == 0:
-        raise ValueError(
-            f"No frames found in {frame_dir} with extensions {frame_extensions}"
-        )
+        raise ValueError(f"No frames found in {frame_dir} with extensions {frame_extensions}")

    frame_files = sorted(frame_files, key=lambda x: x.name)
    total_frames = len(frame_files)
@@ -143,16 +135,11 @@ def _load_video_from_frames(
        frame_indices = list(range(total_frames))
    else:
        if total_frames < num_frames:
-            frame_indices = list(range(total_frames)) + [total_frames - 1] * (
-                num_frames - total_frames)
+            frame_indices = list(range(total_frames)) + [total_frames - 1] * (num_frames - total_frames)
        elif sample_strategy == 'uniform':
-            frame_indices = np.linspace(0,
-                                        total_frames - 1,
-                                        num_frames,
-                                        dtype=int).tolist()
+            frame_indices = np.linspace(0, total_frames - 1, num_frames, dtype=int).tolist()
        elif sample_strategy == 'random':
-            frame_indices = sorted(
-                np.random.choice(total_frames, num_frames, replace=False))
+            frame_indices = sorted(np.random.choice(total_frames, num_frames, replace=False))
        else:
            raise ValueError(f"Unknown sample_strategy: {sample_strategy}")

@@ -170,8 +157,7 @@ def _load_video_from_frames(

    # Stack and convert to tensor
    frames = np.stack(frames)  # [T, H, W, C]
-    frames = torch.from_numpy(frames).permute(0, 3, 1,
-                                              2).float()  # [T, C, H, W]
+    frames = torch.from_numpy(frames).permute(0, 3, 1, 2).float()  # [T, C, H, W]

    return frames

@@ -226,13 +212,12 @@ def load_video_auto(video_path: str | Path,
        raise ValueError(f"Unknown video format at {video_path}")


-def sample_clips_from_video(
-        video: torch.Tensor,
-        num_frames_per_clip: int = 16,
-        num_clips: int = 1,
-        strategy: str | ClipSamplingStrategy = ClipSamplingStrategy.BEGINNING,
-        frame_stride: int = 1,
-        temporal_stride: int = 1) -> list[torch.Tensor]:
+def sample_clips_from_video(video: torch.Tensor,
+                            num_frames_per_clip: int = 16,
+                            num_clips: int = 1,
+                            strategy: str | ClipSamplingStrategy = ClipSamplingStrategy.BEGINNING,
+                            frame_stride: int = 1,
+                            temporal_stride: int = 1) -> list[torch.Tensor]:
    """
    Sample clips from a video with various strategies.
        
@@ -296,10 +281,7 @@ def sample_clips_from_video(
    elif strategy == ClipSamplingStrategy.RANDOM:
        # Sample N random clips
        for _ in range(num_clips):
-            if effective_clip_length == T:
-                start = 0
-            else:
-                start = np.random.randint(0, T - effective_clip_length + 1)
+            start = 0 if effective_clip_length == T else np.random.randint(0, T - effective_clip_length + 1)
            clip = video[start:start + effective_clip_length]
            clips.append(clip)

@@ -312,8 +294,7 @@ def sample_clips_from_video(
            clips.append(clip)
        else:
            # Multiple uniformly spaced clips
-            step = (T - effective_clip_length) / (num_clips -
-                                                  1) if num_clips > 1 else 0
+            step = (T - effective_clip_length) / (num_clips - 1) if num_clips > 1 else 0
            for i in range(num_clips):
                start = int(i * step)
                start = min(start, T - effective_clip_length)
@@ -422,15 +403,12 @@ def load_video_clips_streaming(directory: str | Path,
    failed_count = 0
    total_clips = 0

-    iterator = tqdm(video_paths,
-                    desc="Loading videos") if verbose else video_paths
+    iterator = tqdm(video_paths, desc="Loading videos") if verbose else video_paths

    for video_path in iterator:
        try:
            # Load full video
-            video = load_video_auto(video_path,
-                                    num_frames=None,
-                                    sample_strategy='uniform')
+            video = load_video_auto(video_path, num_frames=None, sample_strategy='uniform')

            # Sample clips from video
            clips = sample_clips_from_video(video,
@@ -445,18 +423,14 @@ def load_video_clips_streaming(directory: str | Path,
                    T, C, H, W = clip.shape
                    if target_size != (H, W):
                        # Resize to target size
-                        clip = clip.contiguous(
-                        )  # Fix non-contiguous tensors first
-                        clip_flat = clip.view(T * C, H,
-                                              W).unsqueeze(0)  # [1, T*C, H, W]
-                        clip_resized = torch.nn.functional.interpolate(
-                            clip_flat,
-                            size=target_size,
-                            mode='bilinear',
-                            align_corners=False)
-                        clip = clip_resized.squeeze(0).view(
-                            T, C, target_size[0],
-                            target_size[1])  # Back to [T, C, H, W]
+                        clip = clip.contiguous()  # Fix non-contiguous tensors first
+                        clip_flat = clip.view(T * C, H, W).unsqueeze(0)  # [1, T*C, H, W]
+                        clip_resized = torch.nn.functional.interpolate(clip_flat,
+                                                                       size=target_size,
+                                                                       mode='bilinear',
+                                                                       align_corners=False)
+                        clip = clip_resized.squeeze(0).view(T, C, target_size[0],
+                                                            target_size[1])  # Back to [T, C, H, W]
                    resized_clips.append(clip)
                clips = resized_clips

@@ -480,11 +454,7 @@ def load_video_clips_streaming(directory: str | Path,

    failure_rate = failed_count / len(video_paths)
    if failure_rate > 0.1:  # More than 10% failed
-        print(
-            f"\nWARNING: {failure_rate:.1%} of videos failed to load ({failed_count}/{len(video_paths)})"
-        )
+        print(f"\nWARNING: {failure_rate:.1%} of videos failed to load ({failed_count}/{len(video_paths)})")

    if verbose:
-        print(
-            f"\nSuccessfully loaded {total_clips} clips from {len(video_paths) - failed_count} videos"
-        )
+        print(f"\nSuccessfully loaded {total_clips} clips from {len(video_paths) - failed_count} videos")
@@ -100,10 +100,7 @@ DEFAULT_PIP_PATTERNS = {
 def run(command):
    """Return (return-code, stdout, stderr)."""
    shell = True if type(command) is str else False
-    p = subprocess.Popen(command,
-                         stdout=subprocess.PIPE,
-                         stderr=subprocess.PIPE,
-                         shell=shell)
+    p = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE, shell=shell)
    raw_output, raw_err = p.communicate()
    rc = p.returncode
    if get_platform() == 'win32':
@@ -157,8 +154,7 @@ def get_conda_packages(run_lambda, patterns=None):
        return out

    return "\n".join(line for line in out.splitlines()
-                     if not line.startswith("#") and any(name in line
-                                                         for name in patterns))
+                     if not line.startswith("#") and any(name in line for name in patterns))


 def get_gcc_version(run_lambda):
@@ -166,27 +162,24 @@ def get_gcc_version(run_lambda):


 def get_clang_version(run_lambda):
-    return run_and_parse_first_match(run_lambda, 'clang --version',
-                                     r'clang version (.*)')
+    return run_and_parse_first_match(run_lambda, 'clang --version', r'clang version (.*)')


 def get_cmake_version(run_lambda):
-    return run_and_parse_first_match(run_lambda, 'cmake --version',
-                                     r'cmake (.*)')
+    return run_and_parse_first_match(run_lambda, 'cmake --version', r'cmake (.*)')


 def get_nvidia_driver_version(run_lambda):
    if get_platform() == 'darwin':
        cmd = 'kextstat | grep -i cuda'
-        return run_and_parse_first_match(run_lambda, cmd,
-                                         r'com[.]nvidia[.]CUDA [(](.*?)[)]')
+        return run_and_parse_first_match(run_lambda, cmd, r'com[.]nvidia[.]CUDA [(](.*?)[)]')
    smi = get_nvidia_smi()
    return run_and_parse_first_match(run_lambda, smi, r'Driver Version: (.*?) ')


 def get_gpu_info(run_lambda):
-    if get_platform() == 'darwin' or (TORCH_AVAILABLE and hasattr(
-            torch.version, 'hip') and torch.version.hip is not None):
+    if get_platform() == 'darwin' or (TORCH_AVAILABLE and hasattr(torch.version, 'hip')
+                                      and torch.version.hip is not None):
        if TORCH_AVAILABLE and torch.cuda.is_available():
            if torch.version.hip is not None:
                prop = torch.cuda.get_device_properties(0)
@@ -208,8 +201,7 @@ def get_gpu_info(run_lambda):


 def get_running_cuda_version(run_lambda):
-    return run_and_parse_first_match(run_lambda, 'nvcc --version',
-                                     r'release .+ V(.*)')
+    return run_and_parse_first_match(run_lambda, 'nvcc --version', r'release .+ V(.*)')


 def get_cudnn_version(run_lambda):
@@ -255,8 +247,7 @@ def get_nvidia_smi():
    if get_platform() == 'win32':
        system_root = os.environ.get('SYSTEMROOT', 'C:\\Windows')
        program_files_root = os.environ.get('PROGRAMFILES', 'C:\\Program Files')
-        legacy_path = os.path.join(program_files_root, 'NVIDIA Corporation',
-                                   'NVSMI', smi)
+        legacy_path = os.path.join(program_files_root, 'NVIDIA Corporation', 'NVSMI', smi)
        new_path = os.path.join(system_root, 'System32', smi)
        smis = [new_path, legacy_path]
        for candidate_smi in smis:
@@ -388,10 +379,8 @@ def get_cpu_info(run_lambda):
    if get_platform() == 'linux':
        rc, out, err = run_lambda('lscpu')
    elif get_platform() == 'win32':
-        rc, out, err = run_lambda(
-            'wmic cpu get Name,Manufacturer,Family,Architecture,ProcessorType,DeviceID, \
-        CurrentClockSpeed,MaxClockSpeed,L2CacheSize,L2CacheSpeed,Revision /VALUE'
-        )
+        rc, out, err = run_lambda('wmic cpu get Name,Manufacturer,Family,Architecture,ProcessorType,DeviceID, \
+        CurrentClockSpeed,MaxClockSpeed,L2CacheSize,L2CacheSpeed,Revision /VALUE')
    elif get_platform() == 'darwin':
        rc, out, err = run_lambda("sysctl -n machdep.cpu.brand_string")
    cpu_info = 'None'
@@ -416,27 +405,22 @@ def get_platform():


 def get_mac_version(run_lambda):
-    return run_and_parse_first_match(run_lambda, 'sw_vers -productVersion',
-                                     r'(.*)')
+    return run_and_parse_first_match(run_lambda, 'sw_vers -productVersion', r'(.*)')


 def get_windows_version(run_lambda):
    system_root = os.environ.get('SYSTEMROOT', 'C:\\Windows')
    wmic_cmd = os.path.join(system_root, 'System32', 'Wbem', 'wmic')
    findstr_cmd = os.path.join(system_root, 'System32', 'findstr')
-    return run_and_read_all(
-        run_lambda,
-        '{} os get Caption | {} /v Caption'.format(wmic_cmd, findstr_cmd))
+    return run_and_read_all(run_lambda, '{} os get Caption | {} /v Caption'.format(wmic_cmd, findstr_cmd))


 def get_lsb_version(run_lambda):
-    return run_and_parse_first_match(run_lambda, 'lsb_release -a',
-                                     r'Description:\t(.*)')
+    return run_and_parse_first_match(run_lambda, 'lsb_release -a', r'Description:\t(.*)')


 def check_release_file(run_lambda):
-    return run_and_parse_first_match(run_lambda, 'cat /etc/*-release',
-                                     r'PRETTY_NAME="(.*)"')
+    return run_and_parse_first_match(run_lambda, 'cat /etc/*-release', r'PRETTY_NAME="(.*)"')


 def get_os(run_lambda):
@@ -499,13 +483,10 @@ def get_pip_packages(run_lambda, patterns=None):
        elif shutil.which("uv") is not None:
            cmd = ["uv", "pip", "list", "--format=freeze"]
        else:
-            raise RuntimeError(
-                "Could not collect pip list output (pip or uv module not available)"
-            )
+            raise RuntimeError("Could not collect pip list output (pip or uv module not available)")

        out = run_and_read_all(run_lambda, cmd)
-        return "\n".join(line for line in out.splitlines()
-                         if any(name in line for name in patterns))
+        return "\n".join(line for line in out.splitlines() if any(name in line for name in patterns))

    pip_version = 'pip3' if sys.version[0] == '3' else 'pip'
    out = run_with_pip()
@@ -537,8 +518,7 @@ def is_xnnpack_available():
 def get_env_vars():
    env_vars = ''
    secret_terms = ('secret', 'token', 'api', 'access', 'password')
-    report_prefix = ("TORCH", "NCCL", "PYTORCH", "CUDA", "CUBLAS", "CUDNN",
-                     "OMP_", "MKL_", "NVIDIA")
+    report_prefix = ("TORCH", "NCCL", "PYTORCH", "CUDA", "CUBLAS", "CUDNN", "OMP_", "MKL_", "NVIDIA")
    for k, v in os.environ.items():
        if any(term in k.lower() for term in secret_terms):
            continue
@@ -559,8 +539,7 @@ def get_env_info():
        debug_mode_str = str(torch.version.debug)
        cuda_available_str = str(torch.cuda.is_available())
        cuda_version_str = torch.version.cuda
-        if not hasattr(torch.version,
-                       'hip') or torch.version.hip is None:  # cuda version
+        if not hasattr(torch.version, 'hip') or torch.version.hip is None:  # cuda version
            hip_compiled_version = hip_runtime_version = miopen_runtime_version = 'N/A'
        else:  # HIP version

@@ -588,9 +567,8 @@ def get_env_info():
    return SystemEnv(
        torch_version=version_str,
        is_debug_build=debug_mode_str,
-        python_version='{} ({}-bit runtime)'.format(
-            sys_version,
-            sys.maxsize.bit_length() + 1),
+        python_version='{} ({}-bit runtime)'.format(sys_version,
+                                                    sys.maxsize.bit_length() + 1),
        python_platform=get_python_platform(),
        is_cuda_available=cuda_available_str,
        cuda_compiled_version=cuda_version_str,
@@ -714,10 +692,8 @@ def pretty_str(envinfo):
        'nvidia_driver_version',
    ]
    all_cuda_fields = dynamic_cuda_fields + ['cudnn_version']
-    all_dynamic_cuda_fields_missing = all(mutable_dict[field] is None
-                                          for field in dynamic_cuda_fields)
-    if TORCH_AVAILABLE and not torch.cuda.is_available(
-    ) and all_dynamic_cuda_fields_missing:
+    all_dynamic_cuda_fields_missing = all(mutable_dict[field] is None for field in dynamic_cuda_fields)
+    if TORCH_AVAILABLE and not torch.cuda.is_available() and all_dynamic_cuda_fields_missing:
        for field in all_cuda_fields:
            mutable_dict[field] = 'No CUDA'
        if envinfo.cuda_compiled_version is None:
@@ -730,19 +706,15 @@ def pretty_str(envinfo):
    mutable_dict = replace_nones(mutable_dict)

    # If either of these are '', replace with 'No relevant packages'
-    mutable_dict['pip_packages'] = replace_if_empty(
-        mutable_dict['pip_packages'])
-    mutable_dict['conda_packages'] = replace_if_empty(
-        mutable_dict['conda_packages'])
+    mutable_dict['pip_packages'] = replace_if_empty(mutable_dict['pip_packages'])
+    mutable_dict['conda_packages'] = replace_if_empty(mutable_dict['conda_packages'])

    # Tag conda and pip packages with a prefix
    # If they were previously None, they'll show up as ie '[conda] Could not collect'
    if mutable_dict['pip_packages']:
-        mutable_dict['pip_packages'] = prepend(
-            mutable_dict['pip_packages'], '[{}] '.format(envinfo.pip_version))
+        mutable_dict['pip_packages'] = prepend(mutable_dict['pip_packages'], '[{}] '.format(envinfo.pip_version))
    if mutable_dict['conda_packages']:
-        mutable_dict['conda_packages'] = prepend(mutable_dict['conda_packages'],
-                                                 '[conda] ')
+        mutable_dict['conda_packages'] = prepend(mutable_dict['conda_packages'], '[conda] ')
    mutable_dict['cpu_info'] = envinfo.cpu_info
    return env_info_fmt.format(**mutable_dict)

@@ -756,18 +728,13 @@ def main():
    output = get_pretty_env_info()
    print(output)

-    if TORCH_AVAILABLE and hasattr(torch, 'utils') and hasattr(
-            torch.utils, '_crash_handler'):
+    if TORCH_AVAILABLE and hasattr(torch, 'utils') and hasattr(torch.utils, '_crash_handler'):
        minidump_dir = torch.utils._crash_handler.DEFAULT_MINIDUMP_DIR
        if sys.platform == "linux" and os.path.exists(minidump_dir):
-            dumps = [
-                os.path.join(minidump_dir, dump)
-                for dump in os.listdir(minidump_dir)
-            ]
+            dumps = [os.path.join(minidump_dir, dump) for dump in os.listdir(minidump_dir)]
            latest = max(dumps, key=os.path.getctime)
            ctime = os.path.getctime(latest)
-            creation_time = datetime.datetime.fromtimestamp(ctime).strftime(
-                '%Y-%m-%d %H:%M:%S')
+            creation_time = datetime.datetime.fromtimestamp(ctime).strftime('%Y-%m-%d %H:%M:%S')
            msg = "\n*** Detected a minidump at {} created on {}, ".format(latest, creation_time) + \
                  "if this is related to your bug please include it when you file a report ***"
            print(msg, file=sys.stderr)
@@ -1,5 +1,4 @@
-from .video_generator.nodes import (NODE_CLASS_MAPPINGS,
-                                    NODE_DISPLAY_NAME_MAPPINGS)
+from .video_generator.nodes import (NODE_CLASS_MAPPINGS, NODE_DISPLAY_NAME_MAPPINGS)

 WEB_DIRECTORY = "./web"
-__all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY']
+__all__ = ['NODE_CLASS_MAPPINGS', 'NODE_DISPLAY_NAME_MAPPINGS', 'WEB_DIRECTORY']
@@ -14,10 +14,7 @@ class LoadImagePath:
    @classmethod
    def INPUT_TYPES(s):
        input_dir = folder_paths.get_input_directory()
-        files = [
-            f for f in os.listdir(input_dir)
-            if os.path.isfile(os.path.join(input_dir, f))
-        ]
+        files = [f for f in os.listdir(input_dir) if os.path.isfile(os.path.join(input_dir, f))]
        files = folder_paths.filter_files_content_types(files, ["image"])
        return {
            "required": {
@@ -65,13 +62,10 @@ class LoadImagePath:
                None,
            ]
            if 'A' in processed_image.getbands():
-                mask = np.array(processed_image.getchannel('A')).astype(
-                    np.float32) / 255.0
+                mask = np.array(processed_image.getchannel('A')).astype(np.float32) / 255.0
                mask = 1. - torch.from_numpy(mask)
            elif processed_image.mode == 'P' and 'transparency' in processed_image.info:
-                mask = np.array(
-                    processed_image.convert('RGBA').getchannel('A')).astype(
-                        np.float32) / 255.0
+                mask = np.array(processed_image.convert('RGBA').getchannel('A')).astype(np.float32) / 255.0
                mask = 1. - torch.from_numpy(mask)
            else:
                mask = torch.zeros((64, 64), dtype=torch.float32, device="cpu")
@@ -9,8 +9,7 @@ from PIL import ImageFile, UnidentifiedImageError
 T = TypeVar('T')


-def conditioning_set_values(conditioning: list[Any],
-                            values: dict[str, Any] | None = None) -> list[Any]:
+def conditioning_set_values(conditioning: list[Any], values: dict[str, Any] | None = None) -> list[Any]:
    if values is None:
        values = {}
    c = []
@@ -27,8 +26,7 @@ def pillow(fn: Callable[[Any], T], arg: Any) -> T:
    prev_value = None
    try:
        x = fn(arg)
-    except (OSError, UnidentifiedImageError, ValueError
-            ):  #PIL issues #4472 and #2445, also fixes ComfyUI issue #3416
+    except (OSError, UnidentifiedImageError, ValueError):  #PIL issues #4472 and #2445, also fixes ComfyUI issue #3416
        prev_value = ImageFile.LOAD_TRUNCATED_IMAGES
        ImageFile.LOAD_TRUNCATED_IMAGES = True
        x = fn(arg)
@@ -39,12 +37,7 @@ def pillow(fn: Callable[[Any], T], arg: Any) -> T:


 def hasher() -> Callable[[], Any]:
-    hashfuncs = {
-        "md5": hashlib.md5,
-        "sha1": hashlib.sha1,
-        "sha256": hashlib.sha256,
-        "sha512": hashlib.sha512
-    }
+    hashfuncs = {"md5": hashlib.md5, "sha1": hashlib.sha1, "sha256": hashlib.sha256, "sha512": hashlib.sha512}
    return hashfuncs[args.default_hashing_function]


@@ -58,8 +51,7 @@ def string_to_torch_dtype(string: str) -> torch.dtype | None:
    return None


-def image_alpha_fix(destination: torch.Tensor,
-                    source: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+def image_alpha_fix(destination: torch.Tensor, source: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
    if destination.shape[-1] < source.shape[-1]:
        source = source[..., :destination.shape[-1]]
    elif destination.shape[-1] > source.shape[-1]:
@@ -26,11 +26,7 @@ class TextEncoderConfig:
    CATEGORY = "fastvideo"

    def set_args(self, prefix, quant_config, lora_config):
-        raw_args = {
-            "prefix": prefix,
-            "quant_config": quant_config,
-            "lora_config": lora_config
-        }
+        raw_args = {"prefix": prefix, "quant_config": quant_config, "lora_config": lora_config}

        # Filter out keys where value is -99999
        args = {k: v for k, v in raw_args.items() if str(int(v)) != str(-99999)}
@@ -13,11 +13,7 @@ from comfy.model_management import processing_interrupted
 from fastvideo import PipelineConfig
 from fastvideo import VideoGenerator as FastVideoGenerator

-sys.path.insert(
-    0,
-    os.path.dirname(
-        os.path.dirname(
-            os.path.dirname(os.path.dirname(os.path.abspath(__file__))))))
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))))


 # Custom exception for interruption
@@ -28,8 +24,7 @@ class GenerationInterruptedException(Exception):
 # Custom exception for interruption that ComfyUI will recognize
 class GenerationCancelledException(Exception):

-    def __init__(self,
-                 message: str = "Generation was cancelled by user") -> None:
+    def __init__(self, message: str = "Generation was cancelled by user") -> None:
        self.message = message
        super().__init__(self.message)

@@ -139,8 +134,7 @@ class VideoGenerator:
                self._generation_interrupted = True

                # Try to send interrupt signal to worker processes
-                if self.generator is not None and hasattr(
-                        self.generator, 'executor'):
+                if self.generator is not None and hasattr(self.generator, 'executor'):
                    try:
                        # The MultiprocExecutor has a workers attribute
                        if hasattr(self.generator.executor, 'workers'):
@@ -156,16 +150,12 @@ class VideoGenerator:
                break
            time.sleep(0.5)

-    def _run_generation(self, prompt: str, output_path: str,
-                        inference_args: dict[str, Any]) -> None:
+    def _run_generation(self, prompt: str, output_path: str, inference_args: dict[str, Any]) -> None:
        """Thread function to run the generation"""
        try:
            if self.generator is not None:
-                self.generator.generate_video(prompt=prompt,
-                                              output_path=output_path,
-                                              **inference_args)
-                self._generation_result = os.path.join(output_path,
-                                                       f"{prompt[:100]}.mp4")
+                self.generator.generate_video(prompt=prompt, output_path=output_path, **inference_args)
+                self._generation_result = os.path.join(output_path, f"{prompt[:100]}.mp4")
            else:
                raise RuntimeError("Generator is not initialized")
        except Exception as e:
@@ -226,8 +216,7 @@ class VideoGenerator:
            update_config_from_args(pipeline_config.vae_config, vae_config)

        if text_encoder_config is not None:
-            update_config_from_args(pipeline_config.text_encoder_configs,
-                                    text_encoder_config)
+            update_config_from_args(pipeline_config.text_encoder_configs, text_encoder_config)

        # Update top-level pipeline config with remaining arguments
        raw_pipeline_args = {}
@@ -245,10 +234,7 @@ class VideoGenerator:
            raw_pipeline_args['text_encoder_precision'] = text_encoder_precision

        # Filter out any value explicitly set to -99999 (auto values)
-        pipeline_args = {
-            k: v
-            for k, v in raw_pipeline_args.items() if str(int(v)) != str(-99999)
-        }
+        pipeline_args = {k: v for k, v in raw_pipeline_args.items() if str(int(v)) != str(-99999)}

        update_config_from_args(pipeline_config, pipeline_args)

@@ -262,38 +248,30 @@ class VideoGenerator:
        if dit_cpu_offload is not None:
            raw_generation_args['dit_cpu_offload'] = dit_cpu_offload

-        generation_args = {
-            k: v
-            for k, v in raw_generation_args.items()
-            if str(int(v)) != str(-99999)
-        }
+        generation_args = {k: v for k, v in raw_generation_args.items() if str(int(v)) != str(-99999)}

        if self.generator is None:
            print('generation_args', generation_args)
            print('pipeline_config', pipeline_config)
-            self.generator = FastVideoGenerator.from_pretrained(
-                model_path=model_path,
-                **generation_args,
-                pipeline_config=pipeline_config)
+            self.generator = FastVideoGenerator.from_pretrained(model_path=model_path,
+                                                                **generation_args,
+                                                                pipeline_config=pipeline_config)

        print('inference_args', inference_args)

        # Start a thread to run the generation
        self._generation_thread = threading.Thread(target=self._run_generation,
-                                                   args=(prompt, output_path,
-                                                         inference_args),
+                                                   args=(prompt, output_path, inference_args),
                                                   daemon=True)
        self._generation_thread.start()

        # Start a background thread to monitor for interruptions
        self._generation_active = True
-        self._interrupt_thread = threading.Thread(
-            target=self._monitor_for_interruption, daemon=True)
+        self._interrupt_thread = threading.Thread(target=self._monitor_for_interruption, daemon=True)
        self._interrupt_thread.start()

        # Wait for either completion or interruption
-        while self._generation_thread.is_alive(
-        ) and not self._interrupt_event.is_set():
+        while self._generation_thread.is_alive() and not self._interrupt_event.is_set():
            self._generation_thread.join(timeout=0.5)

        self._generation_active = False
@@ -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/docs.yml` workflow.
+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.
@@ -0,0 +1,331 @@
+# 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 ──► fix the regression, push, 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 |
+
+If a Full Suite test fails, check the Buildkite build log for the failing step's output.
+Fix the regression, push, and comment `/merge` again to re-trigger.
+
+---
+
+## 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, the developer fixes the issue, pushes, and comments `/merge`
+   again to re-trigger.
+
+**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`) |
+| `/test pre-commit` | Pre-commit checks on PR code | — (runs `ci-precommit.yml` via `workflow_call`) |
+
+**Re-running failed tests:** When you use `/test <name>` to re-run a specific failed test,
+the resulting Buildkite check uses the same name as the original (e.g., `/test encoder`
+creates `buildkite/ci/microscope-encoder-tests`). This overwrites the failed check status.
+Once all tests in a tier pass, the aggregate status (`fastcheck-passed` or
+`full-suite-passed`) is automatically updated to `success` by the `ci-aggregate-status.yml`
+workflow.
+
+**How aggregate status refresh works:**
+
+1. `/test <name>` triggers a Buildkite build with `TEST_SCOPE=direct`. The test step uses
+   the same label as its fastcheck/full-suite counterpart, so the resulting GitHub check
+   overwrites the original.
+2. When the build completes, Buildkite's `notify` posts a `direct-test-completed` commit
+   status. This is the only signal that triggers the aggregate workflow — intermediate step
+   status updates do not trigger it.
+3. `ci-aggregate-status.yml` fires, calls `getCombinedStatusForRef` to fetch the latest
+   status for every context on that commit (each context returns only its most recent
+   state), groups them by prefix (`microscope-*` → fastcheck, `test-tube-*`/`bar-chart-*`
+   → full suite), and posts `fastcheck-passed: success` or `full-suite-passed: success` if
+   all entries in the group are `success`.
+4. Tests that were never triggered (skipped by monorepo-diff) have no status entry and do
+   not block the aggregate.
+
+---
+
+## 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 |
+| `ci-aggregate-status.yml` | Any Buildkite commit status update | Checks if all tests in a tier passed; updates `fastcheck-passed` or `full-suite-passed` |
+| `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 |
@@ -1,65 +1,195 @@
+# RunPod Development Environment

-# 📦 Developing FastVideo on RunPod
+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.

-You can easily use the FastVideo Pod Template on [RunPod](https://www.runpod.io) for development or experimentation.
+## Prerequisites

-## Creating a new pod
+- 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.

- Make sure you are using the correct RunPod account.
 ![RunPod Account Selection](../../assets/images/runpod_account.png)

- Use "Additional Filters" to select CUDA 12.8.
+**2. Filter by CUDA version**
+
+Use "Additional Filters" to select CUDA 12.8.
+
 ![RunPod CUDA selection](../../assets/images/runpod_cuda.png)

- Click "Deploy" and Pick a single A40 or RTX 4090 GPU.
+**3. Select a GPU**
+
+Click "Deploy" and pick a GPU. See [GPU Recommendations](#gpu-recommendations) below for guidance on which GPU to choose.
+
 ![RunPod GPU Selection](../../assets/images/runpod_deploy.png)

- Select the "FastVideo" or "fastvideo-dev" Pod Template.
+**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.
+
 ![RunPod Pod Template Selection](../../assets/images/runpod_create.png)

- Set the Pod name to "`<name>-<FastVideo>-<date>`".
+**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.

- 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)

-## 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:
+Connect with:

 ```bash
-cd /FastVideo
-git pull
+ssh root@<pod-ip> -p <port> -i ~/.ssh/id_ed25519
 ```

-Run your development workflows as usual:
+RunPod also supports VS Code Remote SSH if you prefer an IDE.

-```bash
-# Run linters
-pre-commit run --all-files
+### Custom template (advanced)

-# 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:
+If you're setting up a pod from scratch instead of the FastVideo template, use this image:

 ```
 ghcr.io/hao-ai-lab/fastvideo/fastvideo-dev:py3.12-latest
 ```

-Paste Container Start Command to support SSH ([RunPod Docs](https://docs.runpod.io/pods/configuration/use-ssh)):
+And paste this as the Container Start Command to enable 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.
@@ -0,0 +1,223 @@
+# 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, check the Buildkite build log for the failing step. Fix the
+   issue, push, and comment `/merge` again. You can also re-run individual failed tests
+   with `/test <name>` — see below.
+
+!!! 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
+/test pre-commit       # Pre-commit checks on PR code
+```
+
+The workflow reacts with a 🚀 emoji to confirm the command was received.
+
+When you re-run an individual test with `/test <name>`, the new result overwrites the
+original failed check (same Buildkite check name). Once all tests in a tier pass, the
+`fastcheck-passed` or `full-suite-passed` status is automatically updated.
+
+---
+
+## 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. Check the failing Buildkite 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.
@@ -23,13 +23,18 @@ SSIM tests are located in `fastvideo/tests/ssim`. These tests generate videos us

 ```
 fastvideo/tests/ssim/
-├── <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>
+├── reference_videos/
+│   ├── default/
+│   │   └── <GPU>_reference_videos/
+│   │       ├── <Model_Name>/
+│   │       │   ├── <Backend>/        # e.g., FLASH_ATTN, TORCH_SDPA
+│   │       │   │   └── <Video_File>
+│   └── full_quality/
+│       └── <GPU>_reference_videos/
 ├── test_causal_similarity.py
-├── test_inference_similarity.py
-├── update_reference_videos.sh
+├── test_wan_t2v_similarity.py
+├── test_wan_i2v_similarity.py
+├── reference_videos_cli.py
 └── ...
 ```

@@ -37,7 +42,7 @@ fastvideo/tests/ssim/

 To add a new SSIM test, follow these steps:

-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.
+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.

 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.

@@ -81,10 +86,15 @@ 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`.
+   * 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`.
   * Inspect the generated video to ensure it meets quality expectations.
-   * 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`).
+   * 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`

 ### Running Tests Locally

@@ -96,36 +106,113 @@ pytest fastvideo/tests/ssim/ -vs

 Ensure you have the necessary GPUs available as defined in your test parameters.

-## Modal Workflow
+## CI Integration

-FastVideo uses [Modal](https://modal.com/) for running tests in a CI environment. The workflow scripts are located in `fastvideo/tests/modal/`.
+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.

 ### `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, the `run_ssim_tests` function in `pr_test.py` currently runs:
+For SSIM tests, use `fastvideo/tests/modal/ssim_test.py`:

-```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")
+```bash
+python -m modal run fastvideo/tests/modal/ssim_test.py::run_ssim_tests
 ```

-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.
+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
+```

 ### Workflow Scripts

-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:
+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:
+
 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, you'll need to finally manually update the workflow script in Buildkite. Otherwise, a maintainer will help you update.
+    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.
@@ -0,0 +1,508 @@
+status_definitions:
+  kept: "Public field remains on a public adapter surface with the same meaning."
+  moved: "Public field remains supported but normalizes into a different nested path."
+  profile_owned: "Public field remains supported only through a model/profile-specific surface."
+  compatibility_only: "Legacy public field remains adapter-only during migration and is not part of the canonical typed schema."
+  private_only: "Field should only be handled by private adapters and is not a public FastVideo compatibility promise."
+  internal_only: "Field is runtime/config plumbing and should not be part of the new public typed inference API."
+
+surfaces:
+  fastvideo_args:
+    moved:
+      model_path: generator.model_path
+      workload_type: generator.pipeline.workload_type
+      distributed_executor_backend: generator.engine.execution_backend
+      trust_remote_code: generator.trust_remote_code
+      revision: generator.revision
+      num_gpus: generator.engine.num_gpus
+      tp_size: generator.engine.parallelism.tp_size
+      sp_size: generator.engine.parallelism.sp_size
+      hsdp_replicate_dim: generator.engine.parallelism.hsdp_replicate_dim
+      hsdp_shard_dim: generator.engine.parallelism.hsdp_shard_dim
+      dist_timeout: generator.engine.parallelism.dist_timeout
+      lora_path: generator.pipeline.components.lora_path
+      dit_cpu_offload: generator.engine.offload.dit
+      use_fsdp_inference: generator.engine.use_fsdp_inference
+      dit_layerwise_offload: generator.engine.offload.dit_layerwise
+      text_encoder_cpu_offload: generator.engine.offload.text_encoder
+      image_encoder_cpu_offload: generator.engine.offload.image_encoder
+      vae_cpu_offload: generator.engine.offload.vae
+      pin_cpu_memory: generator.engine.offload.pin_cpu_memory
+      enable_torch_compile: generator.engine.compile.enabled
+      torch_compile_kwargs: generator.engine.compile.kwargs
+      disable_autocast: generator.engine.disable_autocast
+      enable_stage_verification: generator.engine.enable_stage_verification
+      prompt_txt: request.inputs.prompt_path
+      override_text_encoder_safetensors: generator.pipeline.components.text_encoder_weights
+      override_text_encoder_quant: generator.engine.quantization.text_encoder_quant
+      override_transformer_cls_name: generator.pipeline.components.override_transformer_cls_name
+      init_weights_from_safetensors: generator.pipeline.components.transformer_weights
+      init_weights_from_safetensors_2: generator.pipeline.components.transformer_2_weights
+      override_pipeline_cls_name: generator.pipeline.components.override_pipeline_cls_name
+      boundary_ratio: request.sampling.boundary_ratio
+    profile_owned:
+      ltx2_vae_tiling: generator.pipeline.profile_overrides.ltx2.vae_tiling
+      ltx2_vae_spatial_tile_size_in_pixels: generator.pipeline.profile_overrides.ltx2.vae.spatial_tile_size_in_pixels
+      ltx2_vae_spatial_tile_overlap_in_pixels: generator.pipeline.profile_overrides.ltx2.vae.spatial_tile_overlap_in_pixels
+      ltx2_vae_temporal_tile_size_in_frames: generator.pipeline.profile_overrides.ltx2.vae.temporal_tile_size_in_frames
+      ltx2_vae_temporal_tile_overlap_in_frames: generator.pipeline.profile_overrides.ltx2.vae.temporal_tile_overlap_in_frames
+      ltx2_initial_latent_path: request.extensions.ltx2.initial_latent_path
+    compatibility_only:
+      mode: "Legacy multi-mode FastVideoArgs switch; typed inference config should not expose execution mode."
+      inference_mode: "Legacy boolean mirror of mode; kept only through adapters while FastVideoArgs remains."
+      lora_nickname: "Legacy adapter-selection surface pending LoRA API cleanup."
+      lora_target_modules: "Legacy LoRA configuration surface pending dedicated component API."
+      output_type: "Legacy output formatting surface pending GenerationResult cleanup."
+      VSA_sparsity: "Model-specific inference optimization not yet represented in the typed public schema."
+      moba_config_path: "Model-specific MoBA optimization surface not yet represented in the typed public schema."
+      master_port: "Executor/bootstrap compatibility field; not part of the canonical inference schema."
+    private_only:
+      ray_placement_group: "Ray deployment-only field."
+      ray_runtime_env: "Ray deployment-only field."
+    internal_only:
+      pipeline_config: "Legacy internal carrier object."
+      preprocess_config: "Legacy preprocess carrier object."
+      moba_config: "Derived runtime config loaded from moba_config_path."
+      model_paths: "Runtime bookkeeping."
+      model_loaded: "Runtime bookkeeping."
+
+  pipeline_config_base:
+    moved:
+      pipeline_config_path: generator.pipeline.components.pipeline_config_path
+    profile_owned:
+      embedded_cfg_scale: generator.pipeline.profile_overrides.embedded_cfg_scale
+      flow_shift: generator.pipeline.profile_overrides.flow_shift
+      flow_shift_sr: generator.pipeline.profile_overrides.flow_shift_sr
+      is_causal: generator.pipeline.profile_overrides.is_causal
+      vae_tiling: generator.pipeline.profile_overrides.vae_tiling
+      vae_sp: generator.pipeline.profile_overrides.vae_sp
+      dmd_denoising_steps: generator.pipeline.profile_overrides.dmd_denoising_steps
+      ti2v_task: generator.pipeline.profile_overrides.ti2v_task
+      boundary_ratio: generator.pipeline.profile_overrides.boundary_ratio
+    compatibility_only:
+      model_path: "Redundant with generator.model_path."
+      disable_autocast: "Duplicated by generator.engine.disable_autocast during migration."
+      dit_precision: "Precision override pending dedicated typed component precision design."
+      upsampler_precision: "Precision override pending dedicated typed component precision design."
+      vae_precision: "Precision override pending dedicated typed component precision design."
+      image_encoder_precision: "Precision override pending dedicated typed component precision design."
+      text_encoder_precisions: "Precision override pending dedicated typed component precision design."
+    internal_only:
+      dit_config: "Legacy internal component config object."
+      upsampler_config: "Legacy internal component config object."
+      vae_config: "Legacy internal component config object."
+      image_encoder_config: "Legacy internal component config object."
+      text_encoder_configs: "Legacy internal component config object."
+      preprocess_text_funcs: "Internal text preprocessing hooks."
+      postprocess_text_funcs: "Internal text postprocessing hooks."
+
+  pipeline_config_extensions:
+    profile_owned:
+      conditioning_strategy:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      max_num_conditional_frames:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      min_num_conditional_frames:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      sigma_conditional:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      sigma_data:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      state_ch:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      state_t:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      text_encoder_class:
+        sources:
+          - fastvideo.configs.pipelines.cosmos.CosmosConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      autoregressive_chunk_frames:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      autoregressive_overlap_frames:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      cfg_behavior:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      default_camera_rotation:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      default_movement_distance:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      default_negative_prompt:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      default_trajectory_type:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      filter_points_threshold:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      fps:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      frame_buffer_max:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      moge_model_name:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      noise_aug_strength:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      num_frames:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      offload_moge_after_depth:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      use_moge_depth:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      video_resolution:
+        sources:
+          - fastvideo.configs.pipelines.gen3c.Gen3CConfig
+          - fastvideo.configs.pipelines.gen3c.Gen3CInferenceConfig
+      text_encoder_crop_start:
+        sources:
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15I2V480PStepDistilledConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15I2V720PConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15SR1080PConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15T2V480PConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15T2V720PConfig
+          - fastvideo.configs.pipelines.hyworld.HYWorldConfig
+          - fastvideo.configs.pipelines.hyworld.Hunyuan15T2V480PConfig
+      text_encoder_max_lengths:
+        sources:
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15I2V480PStepDistilledConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15I2V720PConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15SR1080PConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15T2V480PConfig
+          - fastvideo.configs.pipelines.hunyuan15.Hunyuan15T2V720PConfig
+          - fastvideo.configs.pipelines.hyworld.HYWorldConfig
+          - fastvideo.configs.pipelines.hyworld.Hunyuan15T2V480PConfig
+      precision:
+        sources:
+          - fastvideo.configs.pipelines.lingbotworld.LingBotWorldI2V480PConfig
+          - fastvideo.configs.pipelines.lingbotworld.Wan2_2_I2V_A14B_Config
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionI2VConfig
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionI2V_A14B_Config
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionT2VConfig
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionT2V_14B_Config
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionT2V_1_3B_Config
+          - fastvideo.configs.pipelines.wan.FastWan2_1_T2V_480P_Config
+          - fastvideo.configs.pipelines.wan.FastWan2_2_TI2V_5B_Config
+          - fastvideo.configs.pipelines.wan.MatrixGameBaseI2V480PConfig
+          - fastvideo.configs.pipelines.wan.MatrixGameI2V480PConfig
+          - fastvideo.configs.pipelines.wan.SelfForcingWan2_2_T2V480PConfig
+          - fastvideo.configs.pipelines.wan.SelfForcingWanT2V480PConfig
+          - fastvideo.configs.pipelines.wan.WANV2VConfig
+          - fastvideo.configs.pipelines.wan.Wan2_2_I2V_A14B_Config
+          - fastvideo.configs.pipelines.wan.Wan2_2_T2V_A14B_Config
+          - fastvideo.configs.pipelines.wan.Wan2_2_TI2V_5B_Config
+          - fastvideo.configs.pipelines.wan.WanI2V480PConfig
+          - fastvideo.configs.pipelines.wan.WanI2V720PConfig
+          - fastvideo.configs.pipelines.wan.WanT2V480PConfig
+          - fastvideo.configs.pipelines.wan.WanT2V720PConfig
+      warp_denoising_step:
+        sources:
+          - fastvideo.configs.pipelines.lingbotworld.LingBotWorldI2V480PConfig
+          - fastvideo.configs.pipelines.lingbotworld.Wan2_2_I2V_A14B_Config
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionI2VConfig
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionI2V_A14B_Config
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionT2VConfig
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionT2V_14B_Config
+          - fastvideo.configs.pipelines.turbodiffusion.TurboDiffusionT2V_1_3B_Config
+          - fastvideo.configs.pipelines.wan.FastWan2_1_T2V_480P_Config
+          - fastvideo.configs.pipelines.wan.FastWan2_2_TI2V_5B_Config
+          - fastvideo.configs.pipelines.wan.MatrixGameBaseI2V480PConfig
+          - fastvideo.configs.pipelines.wan.MatrixGameI2V480PConfig
+          - fastvideo.configs.pipelines.wan.SelfForcingWan2_2_T2V480PConfig
+          - fastvideo.configs.pipelines.wan.SelfForcingWanT2V480PConfig
+          - fastvideo.configs.pipelines.wan.WANV2VConfig
+          - fastvideo.configs.pipelines.wan.Wan2_2_I2V_A14B_Config
+          - fastvideo.configs.pipelines.wan.Wan2_2_T2V_A14B_Config
+          - fastvideo.configs.pipelines.wan.Wan2_2_TI2V_5B_Config
+          - fastvideo.configs.pipelines.wan.WanI2V480PConfig
+          - fastvideo.configs.pipelines.wan.WanI2V720PConfig
+          - fastvideo.configs.pipelines.wan.WanT2V480PConfig
+          - fastvideo.configs.pipelines.wan.WanT2V720PConfig
+      bsa_cdf_threshold:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      bsa_chunk_k:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      bsa_chunk_q:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      bsa_params:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      bsa_sparsity:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      enable_bsa:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      enable_kv_cache:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      enhance_hf:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      offload_kv_cache:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      t_thresh:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      use_distill:
+        sources: [fastvideo.configs.pipelines.longcat.LongCatT2V480PConfig, fastvideo.configs.pipelines.longcat.LongCatT2V704PConfig]
+      scheduler_arch:
+        sources: [fastvideo.configs.pipelines.sd35.SD35Config]
+      text_encoder_archs:
+        sources: [fastvideo.configs.pipelines.sd35.SD35Config]
+      tokenizer_archs:
+        sources: [fastvideo.configs.pipelines.sd35.SD35Config]
+      transformer_arch:
+        sources: [fastvideo.configs.pipelines.sd35.SD35Config]
+      vae_arch:
+        sources: [fastvideo.configs.pipelines.sd35.SD35Config]
+      expand_timesteps:
+        sources:
+          - fastvideo.configs.pipelines.wan.FastWan2_2_TI2V_5B_Config
+          - fastvideo.configs.pipelines.wan.Wan2_2_TI2V_5B_Config
+      context_noise:
+        sources: [fastvideo.configs.pipelines.wan.MatrixGameI2V480PConfig]
+      num_frames_per_block:
+        sources: [fastvideo.configs.pipelines.wan.MatrixGameI2V480PConfig]
+    compatibility_only:
+      batch_size: "Gen3C inference-only tuning field pending typed batching design."
+      gradient_checkpointing: "Gen3C inference-only compatibility field pending typed batching design."
+      guidance_scale: "Gen3C pipeline-level default pending profile/default-request cleanup."
+      num_inference_steps: "Gen3C pipeline-level default pending profile/default-request cleanup."
+    internal_only:
+      audio_decoder_config: "Legacy internal component config object."
+      audio_decoder_precision: "Precision override pending dedicated component precision design."
+      vocoder_config: "Legacy internal component config object."
+      vocoder_precision: "Precision override pending dedicated component precision design."
+
+  sampling_param_base:
+    moved:
+      image_path: request.inputs.image_path
+      pil_image: request.inputs.pil_image
+      video_path: request.inputs.video_path
+      mouse_cond: request.inputs.mouse_cond
+      keyboard_cond: request.inputs.keyboard_cond
+      grid_sizes: request.inputs.grid_sizes
+      pose: request.inputs.pose
+      c2ws_plucker_emb: request.inputs.c2ws_plucker_emb
+      refine_from: request.inputs.refine_from
+      stage1_video: request.inputs.stage1_video
+      prompt: request.prompt
+      negative_prompt: request.negative_prompt
+      prompt_path: request.inputs.prompt_path
+      output_path: request.output.output_path
+      output_video_name: request.output.output_video_name
+      num_videos_per_prompt: request.sampling.num_videos_per_prompt
+      seed: request.sampling.seed
+      num_frames: request.sampling.num_frames
+      height: request.sampling.height
+      width: request.sampling.width
+      height_sr: request.sampling.height_sr
+      width_sr: request.sampling.width_sr
+      fps: request.sampling.fps
+      num_inference_steps: request.sampling.num_inference_steps
+      num_inference_steps_sr: request.sampling.num_inference_steps_sr
+      guidance_scale: request.sampling.guidance_scale
+      guidance_rescale: request.sampling.guidance_rescale
+      boundary_ratio: request.sampling.boundary_ratio
+      sigmas: request.sampling.sigmas
+      enable_teacache: request.runtime.enable_teacache
+      save_video: request.output.save_video
+      return_frames: request.output.return_frames
+      return_trajectory_latents: request.runtime.return_trajectory_latents
+      return_trajectory_decoded: request.runtime.return_trajectory_decoded
+    profile_owned:
+      t_thresh: request.stage_overrides.refine.t_thresh
+      spatial_refine_only: request.stage_overrides.refine.spatial_refine_only
+      num_cond_frames: request.stage_overrides.refine.num_cond_frames
+      trajectory_type: request.extensions.gen3c.trajectory_type
+      movement_distance: request.extensions.gen3c.movement_distance
+      camera_rotation: request.extensions.gen3c.camera_rotation
+    internal_only:
+      data_type: "Derived from the request shape and not a public input."
+
+  sampling_param_extensions:
+    moved:
+      guidance_scale_2:
+        target: request.sampling.guidance_scale_2
+        sources:
+          - fastvideo.configs.sample.lingbotworld.LingBotWorld_SamplingParam
+          - fastvideo.configs.sample.lingbotworld.Wan2_2_I2V_A14B_SamplingParam
+          - fastvideo.configs.sample.wan.SelfForcingWan2_2_T2V_A14B_480P_SamplingParam
+          - fastvideo.configs.sample.wan.Wan2_2_I2V_A14B_SamplingParam
+          - fastvideo.configs.sample.wan.Wan2_2_T2V_A14B_SamplingParam
+    profile_owned:
+      action_list:
+        target: request.extensions.hunyuangamecraft.action_list
+        sources:
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraftSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft65FrameSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft129FrameSamplingParam
+      action_speed_list:
+        target: request.extensions.hunyuangamecraft.action_speed_list
+        sources:
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraftSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft65FrameSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft129FrameSamplingParam
+      camera_states:
+        target: request.extensions.hunyuangamecraft.camera_states
+        sources:
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraftSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft65FrameSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft129FrameSamplingParam
+      camera_trajectory:
+        target: request.extensions.hunyuangamecraft.camera_trajectory
+        sources:
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraftSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft65FrameSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft129FrameSamplingParam
+      conditioning_mask:
+        target: request.extensions.hunyuangamecraft.conditioning_mask
+        sources:
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraftSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft65FrameSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft129FrameSamplingParam
+      gt_latents:
+        target: request.extensions.hunyuangamecraft.gt_latents
+        sources:
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraftSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft65FrameSamplingParam
+          - fastvideo.configs.sample.hunyuangamecraft.HunyuanGameCraft129FrameSamplingParam
+      prompt_attention_mask:
+        target: request.extensions.hyworld.prompt_attention_mask
+        sources: [fastvideo.configs.sample.hyworld.HYWorld_SamplingParam]
+      negative_attention_mask:
+        target: request.extensions.hyworld.negative_attention_mask
+        sources: [fastvideo.configs.sample.hyworld.HYWorld_SamplingParam]
+      ltx2_cfg_scale_audio:
+        target: request.extensions.ltx2.cfg_scale_audio
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_cfg_scale_video:
+        target: request.extensions.ltx2.cfg_scale_video
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_modality_scale_audio:
+        target: request.extensions.ltx2.modality_scale_audio
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_modality_scale_video:
+        target: request.extensions.ltx2.modality_scale_video
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_rescale_scale:
+        target: request.extensions.ltx2.rescale_scale
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_stg_blocks_audio:
+        target: request.extensions.ltx2.stg_blocks_audio
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_stg_blocks_video:
+        target: request.extensions.ltx2.stg_blocks_video
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_stg_scale_audio:
+        target: request.extensions.ltx2.stg_scale_audio
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+      ltx2_stg_scale_video:
+        target: request.extensions.ltx2.stg_scale_video
+        sources: [fastvideo.configs.sample.ltx2.LTX2BaseSamplingParam]
+
+  openai_image_request:
+    kept:
+      model: "HTTP adapter model-routing field."
+      response_format: "HTTP adapter response formatting field."
+      output_format: "HTTP adapter output-format field."
+      background: "HTTP adapter output-format field."
+      quality: "Compatibility field currently accepted by the adapter."
+      style: "Compatibility field currently accepted by the adapter."
+      user: "Compatibility field currently accepted by the adapter."
+    moved:
+      prompt: request.prompt
+      n: request.sampling.num_videos_per_prompt
+      size:
+        target: request.sampling.width,height
+        note: "Adapter parses OpenAI size strings as WIDTHxHEIGHT and forwards width then height."
+      num_inference_steps: request.sampling.num_inference_steps
+      guidance_scale: request.sampling.guidance_scale
+      true_cfg_scale: request.sampling.true_cfg_scale
+      seed: request.sampling.seed
+      negative_prompt: request.negative_prompt
+      enable_teacache: request.runtime.enable_teacache
+
+  openai_video_request:
+    kept:
+      model: "HTTP adapter model-routing field."
+    moved:
+      prompt: request.prompt
+      input_reference: request.inputs.image_path
+      reference_url: request.inputs.image_path
+      size:
+        target: request.sampling.width,height
+        note: "Adapter parses OpenAI size strings as WIDTHxHEIGHT and forwards width then height."
+      fps: request.sampling.fps
+      num_frames: request.sampling.num_frames
+      seed: request.sampling.seed
+      num_inference_steps: request.sampling.num_inference_steps
+      guidance_scale: request.sampling.guidance_scale
+      guidance_scale_2: request.sampling.guidance_scale_2
+      true_cfg_scale: request.sampling.true_cfg_scale
+      negative_prompt: request.negative_prompt
+      enable_teacache: request.runtime.enable_teacache
+      output_path: request.output.output_path
+    compatibility_only:
+      seconds:
+        target: request.sampling.num_frames
+        note: "HTTP adapter duration convenience field. If num_frames is omitted, the adapter computes num_frames = fps * seconds."
+
+cli:
+  notes:
+    - "CLI parity is checked against the actual generate/serve parser dest sets."
+    - "The inventory tracks parser dest names, excluding argparse's implicit help action."
+    - "The refactored inference CLI is config-only: subcommands expose only --config, and any additional CLI input must use dotted override paths."
+  generate:
+    explicit_local_fields:
+      - config
+    expected_dests:
+      - config
+  serve:
+    explicit_local_fields:
+      - config
+    expected_dests:
+      - config
@@ -16,7 +16,8 @@ Both models are trained on **61×448×832** resolution but support generating vi
 First install [VSA](../attention/vsa/index.md). Set `MODEL_BASE` to your own model path and run:

 ```bash
-bash scripts/inference/v1_inference_wan_dmd.sh
+FASTVIDEO_ATTENTION_BACKEND=VIDEO_SPARSE_ATTN \
+  fastvideo generate --config scripts/inference/inference_wan_VSA_DMD_1_3B.yaml
 ```

 ## 🗂️ Dataset
@@ -42,8 +42,7 @@ def fix_case(text: str) -> str:
        r"int\d+": lambda x: x.group(0).upper(),  # e.g. int8, int16
    }
    for pattern, repl in subs.items():
-        text = re.sub(rf'\b{pattern}\b', repl, text,
-                      flags=re.IGNORECASE)  # type: ignore[call-overload]
+        text = re.sub(rf'\b{pattern}\b', repl, text, flags=re.IGNORECASE)  # type: ignore[call-overload]
    return text


@@ -134,9 +133,7 @@ class Example:
        if not markdown_files:
            raise IndexError(f"No Markdown files found in {self.path}")

-        readme_files = [
-            f for f in markdown_files if f.name.lower() == "readme.md"
-        ]
+        readme_files = [f for f in markdown_files if f.name.lower() == "readme.md"]
        if readme_files:
            return readme_files[0]

@@ -156,8 +153,7 @@ class Example:
        if self.path.is_file():
            return []
        is_other_file = lambda file: file.is_file() and file != self.main_file
-        return [file for file in self.path.rglob("*")
-                if is_other_file(file)]  # type: ignore[no-untyped-call]
+        return [file for file in self.path.rglob("*") if is_other_file(file)]  # type: ignore[no-untyped-call]

    def determine_title(self) -> str:
        return fix_case(self.path.stem.replace("_", " ").title())
@@ -190,8 +186,8 @@ class Example:
        content += "## Additional Files\n\n"
        # Define binary/non-text file extensions to skip
        binary_extensions = {
-            '.mp4', '.avi', '.mov', '.mkv', '.gif', '.jpg', '.jpeg', '.png',
-            '.webp', '.bmp', '.pdf', '.zip', '.tar', '.gz', '.mp3', '.wav'
+            '.mp4', '.avi', '.mov', '.mkv', '.gif', '.jpg', '.jpeg', '.png', '.webp', '.bmp', '.pdf', '.zip', '.tar',
+            '.gz', '.mp3', '.wav'
        }

        for file in sorted(self.other_files):
@@ -259,8 +255,7 @@ def create_category_indices() -> dict[str, Index]:
    category_indices = {
        "inference":
        Index(
-            path=ROOT_DIR /
-            "docs/inference/examples/examples_inference_index.md",
+            path=ROOT_DIR / "docs/inference/examples/examples_inference_index.md",
            title="🚀 Examples",
            description=
            "Inference examples demonstrate how to use FastVideo inference. We recommend starting with [basic.md](basic.md).",
@@ -271,18 +266,15 @@ def create_category_indices() -> dict[str, Index]:
        Index(
            path=ROOT_DIR / "docs/training/examples/examples_training_index.md",
            title="🚀 Examples",
-            description=
-            "Training examples demonstrate how to use FastVideo training.",
+            description="Training examples demonstrate how to use FastVideo training.",
            caption="Examples",
            maxdepth=3,
        ),
        "distillation":
        Index(
-            path=ROOT_DIR /
-            "docs/distillation/examples/examples_distillation_index.md",
+            path=ROOT_DIR / "docs/distillation/examples/examples_distillation_index.md",
            title="🚀 Examples",
-            description=
-            "Distillation examples demonstrate how to use FastVideo distillation.",
+            description="Distillation examples demonstrate how to use FastVideo distillation.",
            caption="Examples",
            maxdepth=3,
        ),
@@ -296,8 +288,7 @@ def create_category_indices() -> dict[str, Index]:
    return category_indices


-def find_examples(category_indices: dict[str, Index],
-                  generate_main_index: bool) -> list[Example]:
+def find_examples(category_indices: dict[str, Index], generate_main_index: bool) -> list[Example]:
    """Find all examples from the examples directory."""
    examples = []
    glob_patterns = ["*.py", "*.md", "*.sh"]
@@ -339,13 +330,9 @@ def find_examples(category_indices: dict[str, Index],
    return examples


-def create_nested_structures(
-    examples: list[Example]
-) -> dict[str, dict[str, dict[str, dict[str, NestedStructure]]]]:
+def create_nested_structures(examples: list[Example]) -> dict[str, dict[str, dict[str, dict[str, NestedStructure]]]]:
    """Create nested structures for training and distillation categories."""
-    nested_structures: dict[str, dict[str, dict[str,
-                                                dict[str,
-                                                     NestedStructure]]]] = {}
+    nested_structures: dict[str, dict[str, dict[str, dict[str, NestedStructure]]]] = {}

    # Map category names to actual directory names
    category_dir_mapping = {
@@ -378,13 +365,11 @@ def create_nested_structures(
                    nested_structures[example.category][method][model] = {}

                # Store the nested structure
-                nested_structures[
-                    example.category][method][model][dataset] = NestedStructure(
-                        category=example.category,
-                        method=method,
-                        model=model,
-                        dataset=dataset,
-                        example=example)
+                nested_structures[example.category][method][model][dataset] = NestedStructure(category=example.category,
+                                                                                              method=method,
+                                                                                              model=model,
+                                                                                              dataset=dataset,
+                                                                                              example=example)

        elif example.category == "distillation" and len(path_parts) >= 2:
            # For distillation examples like Wan2.1-T2V/Wan-Syn-Data-480P
@@ -401,20 +386,16 @@ def create_nested_structures(
                nested_structures[example.category][method][model] = {}

            # Store the nested structure
-            nested_structures[
-                example.category][method][model][dataset] = NestedStructure(
-                    category=example.category,
-                    method=method,
-                    model=model,
-                    dataset=dataset,
-                    example=example)
+            nested_structures[example.category][method][model][dataset] = NestedStructure(category=example.category,
+                                                                                          method=method,
+                                                                                          model=model,
+                                                                                          dataset=dataset,
+                                                                                          example=example)

    return nested_structures


-def generate_flat_examples(examples: list[Example],
-                           category_indices: dict[str, Index],
-                           examples_index: Index | None,
+def generate_flat_examples(examples: list[Example], category_indices: dict[str, Index], examples_index: Index | None,
                           generate_main_index: bool) -> None:
    """Generate documentation for flat structure examples (inference, etc.)."""
    for example in examples:
@@ -437,9 +418,8 @@ def generate_flat_examples(examples: list[Example],
        index.documents.append(example.path.stem)


-def generate_nested_examples(nested_structures: dict[str, dict[str, dict[
-    str, dict[str, NestedStructure]]]], category_indices: dict[str,
-                                                               Index]) -> None:
+def generate_nested_examples(nested_structures: dict[str, dict[str, dict[str, dict[str, NestedStructure]]]],
+                             category_indices: dict[str, Index]) -> None:
    """Generate documentation for nested structure examples (training, distillation)."""
    for category_name in ["training", "distillation"]:
        if category_name not in category_indices or category_name not in nested_structures:
@@ -464,12 +444,11 @@ def generate_nested_examples(nested_structures: dict[str, dict[str, dict[
                        f.write(nested_struct.example.generate())

                # Create model-level index
-                model_index = Index(
-                    path=category_base_dir / f"{model}.md",
-                    title=fix_case(model.replace('_', ' ')),
-                    description=f"Examples for the {model} model.",
-                    caption=f"{fix_case(model.replace('_', ' '))} Datasets",
-                    maxdepth=1)
+                model_index = Index(path=category_base_dir / f"{model}.md",
+                                    title=fix_case(model.replace('_', ' ')),
+                                    description=f"Examples for the {model} model.",
+                                    caption=f"{fix_case(model.replace('_', ' '))} Datasets",
+                                    maxdepth=1)

                # Add dataset indices to model index
                for dataset, nested_struct in datasets.items():
@@ -508,8 +487,7 @@ def generate_examples(generate_main_index: bool = False) -> None:
        examples_index = Index(
            path=main_index_dir / "examples_index.md",
            title="💡 Examples",
-            description=
-            "A collection of examples demonstrating usage of FastVideo.\n\n"
+            description="A collection of examples demonstrating usage of FastVideo.\n\n"
            f"All documented examples are autogenerated using [generate_examples.py](https://github.com/{GITHUB_REPO}/blob/main/docs/generate_examples.py) "
            f"from examples found in the [examples](https://github.com/{GITHUB_REPO}/tree/main/examples) directory.",
            caption="Examples",
@@ -522,8 +500,7 @@ def generate_examples(generate_main_index: bool = False) -> None:
    nested_structures = create_nested_structures(examples)

    # Generate flat structure examples (inference, etc.)
-    generate_flat_examples(examples, category_indices, examples_index,
-                           generate_main_index)
+    generate_flat_examples(examples, category_indices, examples_index, generate_main_index)

    # Generate nested structure examples (training, distillation)
    generate_nested_examples(nested_structures, category_indices)
@@ -534,11 +511,8 @@ def generate_examples(generate_main_index: bool = False) -> None:
            # Add to main index if it exists
            if generate_main_index and examples_index:
                main_index_dir = examples_index.path.parent
-                rel_path = os.path.relpath(category_index.path,
-                                           start=main_index_dir)
-                examples_index.documents.insert(
-                    0,
-                    str(rel_path).replace("\\", "/").replace(".md", ""))
+                rel_path = os.path.relpath(category_index.path, start=main_index_dir)
+                examples_index.documents.insert(0, str(rel_path).replace("\\", "/").replace(".md", ""))

            # Write the category index file
            with open(category_index.path, "w+") as f:
@@ -455,6 +455,6 @@ User: generator.generate_video(prompt, ...)
   `fastvideo/pipelines/stages/`, implement `forward()`, optionally
   implement `verify_input()`/`verify_output()`.

-7. **Verify** — Run `fastvideo generate --model-path <path> --prompt
-   "test" --num-inference-steps 2` to confirm the pipeline loads and
-   generates output.
+7. **Verify** — Run `fastvideo generate --config <config.yaml>` with a
+   minimal nested config to confirm the pipeline loads and generates
+   output.
@@ -1,71 +1,29 @@
 # FastVideo CLI Inference

-The FastVideo CLI exposes the same core inference controls as the Python API.
+The FastVideo CLI is config-first. Inference runs are driven by a nested JSON or
+YAML config, with optional dotted-path overrides on the command line. The
+contract matches training: use an explicit subcommand plus `--config`, then add
+any dotted overrides you need.

 ## Basic Usage

-Use either:
-
-1. `--model-path` + `--prompt`
-2. `--model-path` + `--prompt-txt` (batch prompts, one line per prompt)
-3. `--config` (JSON/YAML)
-
 ```bash
-fastvideo generate --model-path Wan-AI/Wan2.1-T2V-1.3B-Diffusers \
-  --prompt "A cat playing with a ball of yarn"
+fastvideo generate --config config.yaml
+fastvideo serve --config serve.yaml
 ```

-```bash
-fastvideo generate --model-path Wan-AI/Wan2.1-T2V-1.3B-Diffusers \
-  --prompt-txt prompts.txt
-```
-
-You cannot provide both `--prompt` and `--prompt-txt` in the same run.
-
 ## View All Arguments

 ```bash
 fastvideo generate --help
 ```

-Arguments come from:
+The subcommands intentionally expose only `--config`. Any per-run CLI changes
+must use dotted override paths such as:

- FastVideo runtime args (`FastVideoArgs`)
- Sampling args (`SamplingParam`)
- Pipeline config args (`PipelineConfig`)
-
-## Common Arguments
-
-### Parallelism
-
- `--num-gpus`
- `--sp-size`
- `--tp-size`
-
-### Sampling
-
- `--num-frames`
- `--height` / `--width`
- `--num-inference-steps`
- `--guidance-scale`
- `--seed`
- `--negative-prompt`
-
-### Output
-
- `--output-path`
- `--save-video` / `--no-save-video`
- `--return-frames`
-
-### Offloading and Performance
-
- `--dit-layerwise-offload`
- `--use-fsdp-inference`
- `--text-encoder-cpu-offload`
- `--image-encoder-cpu-offload`
- `--vae-cpu-offload`
- `--enable-torch-compile`
- `--torch-compile-kwargs`
+- `--generator.engine.num_gpus 2`
+- `--request.sampling.seed 42`
+- `--server.port 9000`

 ## Using Config Files

@@ -73,50 +31,53 @@ Arguments come from:
 fastvideo generate --config config.yaml
 ```

-Config files can be JSON or YAML. CLI flags override config-file values.
+Config files can be JSON or YAML. Dotted CLI overrides take precedence over
+config-file values.

 Example `config.yaml`:

 ```yaml
-model_path: "FastVideo/FastHunyuan-diffusers"
-prompt: "A capybara lounging in a hammock"
-output_path: "outputs/"
-num_gpus: 2
-sp_size: 2
-tp_size: 1
-num_frames: 45
-height: 720
-width: 1280
-num_inference_steps: 6
-seed: 1024
-dit_precision: "bf16"
-vae_precision: "fp16"
-vae_tiling: true
-vae_sp: true
-enable_torch_compile: false
+generator:
+  model_path: FastVideo/FastHunyuan-diffusers
+  engine:
+    num_gpus: 2
+    parallelism:
+      sp_size: 2
+      tp_size: 1
+request:
+  prompt: A capybara lounging in a hammock
+  sampling:
+    num_frames: 45
+    height: 720
+    width: 1280
+    num_inference_steps: 6
+    seed: 1024
+  output:
+    output_path: outputs/
 ```

 Notes:

- Use `dit_precision` / `vae_precision` (not `precision`).
- Nested config objects are supported, for example `vae_config` and
-  `dit_config`.
+- `generator` and `request` are the top-level keys for generation configs.
+- `serve` configs use `generator`, `server`, and optional `default_request`.
+- Prompt text files belong under `request.inputs.prompt_path`.

 ## Examples

 Simple generation:

 ```bash
-fastvideo generate \
-  --model-path FastVideo/FastHunyuan-diffusers \
-  --prompt "A cat playing with a ball of yarn" \
-  --num-frames 45 --height 720 --width 1280 \
-  --num-inference-steps 6 --seed 1024 \
-  --output-path outputs/
+fastvideo generate --config config.yaml
 ```

-Config + CLI override:
+Config + dotted override:

 ```bash
-fastvideo generate --config config.yaml --prompt "A panda skiing at sunset"
+fastvideo generate --config config.yaml --request.prompt "A panda skiing at sunset"
+```
+
+Helper wrapper with positional config path:
+
+```bash
+bash scripts/inference/run.sh scripts/inference/inference_wan.yaml
 ```
@@ -73,32 +73,40 @@ if __name__ == '__main__':

 ## JSON/YAML Config Files (CLI)

-The CLI supports `--config` with JSON or YAML. Command-line arguments override
-config file values.
-By default, `fastvideo generate` uses `return_frames=false` unless you set
-`--return-frames` (or `return_frames: true` in config).
+The inference CLI is config-first. Use an explicit subcommand with `--config`,
+then apply optional dotted overrides on top, matching the training CLI style.
+By default, CLI generation uses `return_frames=false` unless you set
+`request.output.return_frames: true` in config or via a dotted override.

 ```bash
 fastvideo generate --config config.yaml
 ```

-Use CLI argument names as keys (underscore or hyphen is accepted). Example:
+Example nested config:

 ```yaml
-model_path: "FastVideo/FastHunyuan-diffusers"
-prompt: "A capybara relaxing in a hammock"
-num_gpus: 2
-sp_size: 2
-num_frames: 45
-height: 720
-width: 1280
-num_inference_steps: 6
-seed: 1024
-dit_precision: "bf16"
-vae_precision: "fp16"
-vae_tiling: true
-vae_sp: true
-enable_torch_compile: false
+generator:
+  model_path: FastVideo/FastHunyuan-diffusers
+  engine:
+    num_gpus: 2
+    parallelism:
+      sp_size: 2
+request:
+  prompt: A capybara relaxing in a hammock
+  sampling:
+    num_frames: 45
+    height: 720
+    width: 1280
+    num_inference_steps: 6
+    seed: 1024
+  output:
+    output_path: outputs/
+```
+
+Override individual values from the CLI with dotted paths:
+
+```bash
+fastvideo generate --config config.yaml --request.sampling.seed 42
 ```

 ## Performance Optimization
@@ -0,0 +1,129 @@
+# GEN3C: 3D-Informed Camera-Controlled Video Generation
+
+[GEN3C](https://arxiv.org/abs/2503.03751) is NVIDIA's Cosmos-7B-based video model for camera-controlled generation from a single image. The FastVideo integration supports the GEN3C I2V workflow, including 3D cache conditioning and tokenizer-based conditioning latents.
+
+## Key Features
+
+- **Camera trajectory control**: `left/right/up/down/zoom_in/zoom_out/clockwise/counterclockwise`
+- **3D cache conditioning**: depth prediction -> point cloud cache -> forward warping -> latent conditioning
+- **Single-image to video generation**: 121-frame generation with camera motion
+- **Official raw checkpoint conversion**: `model.pt` -> Diffusers/FastVideo layout
+
+## Model Sources
+
+- Official raw checkpoint (not Diffusers): `nvidia/GEN3C-Cosmos-7B`
+- Diffusers-format checkpoint: `FastVideo/GEN3C-Cosmos-7B-Diffusers`
+
+## Prerequisites
+
+- Install MoGe:
+
+```bash
+pip install git+https://github.com/microsoft/MoGe.git
+```
+
+- If you hit `ImportError: libGL.so.1` (common on Ubuntu/headless nodes), you can try installing OpenCV runtime libs:
+
+```bash
+sudo apt-get update
+sudo apt-get install -y libgl1 libglib2.0-0 libsm6 libxext6 libxrender1
+```
+
+## Quick Start
+
+### Option A: Use Diffusers-format weights directly
+
+```bash
+python examples/inference/basic/basic_gen3c.py \
+  --model_path FastVideo/GEN3C-Cosmos-7B-Diffusers \
+  --image_path /path/to/input.png \
+  --prompt "" \
+  --trajectory left \
+  --movement_distance 0.3 \
+  --camera_rotation center_facing \
+  --num_inference_steps 35 \
+  --guidance_scale 1.0 \
+  --output_path outputs_video/gen3c_output.mp4
+```
+
+### Option B: Convert official raw checkpoint locally
+
+1. Download:
+
+```bash
+huggingface-cli download nvidia/GEN3C-Cosmos-7B --local-dir official_weights/GEN3C-Cosmos-7B
+```
+
+1. Convert:
+
+```bash
+python scripts/checkpoint_conversion/convert_gen3c_to_fastvideo.py \
+  --source official_weights/GEN3C-Cosmos-7B/model.pt \
+  --output converted_weights/GEN3C-Cosmos-7B
+```
+
+1. Run:
+
+```bash
+python examples/inference/basic/basic_gen3c.py \
+  --model_path converted_weights/GEN3C-Cosmos-7B \
+  --image_path /path/to/input.png \
+  --prompt "" \
+  --trajectory left \
+  --movement_distance 0.3 \
+  --camera_rotation center_facing \
+  --num_inference_steps 35 \
+  --guidance_scale 1.0 \
+  --output_path outputs_video/gen3c_output.mp4
+```
+
+## FastVideo Defaults
+
+GEN3C defaults in FastVideo:
+
+- `height=704`, `width=1280`
+- `num_frames=121`
+- `num_inference_steps=35`
+- `guidance_scale=1.0`
+- `fps=24`
+
+These values are defined in:
+
+- `fastvideo/configs/sample/gen3c.py`
+- `fastvideo/configs/pipelines/gen3c.py`
+
+and align with the official GEN3C inference defaults in:
+
+- `tmp/GEN3C/cosmos_predict1/diffusion/inference/inference_utils.py`
+
+## Scheduler Note
+
+The converted GEN3C Diffusers layout may include a FlowMatch scheduler config, but GEN3C denoising uses EDM preconditioning behavior. FastVideo's GEN3C pipeline enforces an EDM scheduler at runtime for parity with official inference behavior.
+
+Implementation path:
+
+- `fastvideo/pipelines/basic/gen3c/gen3c_pipeline.py`
+
+## 3D Cache Conditioning Path
+
+FastVideo GEN3C conditioning stage performs:
+
+1. MoGe depth estimation from input image
+2. 3D cache initialization
+3. Camera trajectory generation
+4. Forward rendering of warped frames + masks
+5. VAE/tokenizer encoding of conditioning buffers
+6. Denoising with condition mask + condition pose channels
+
+Main implementation:
+
+- `fastvideo/pipelines/basic/gen3c/gen3c_pipeline.py`
+- `fastvideo/pipelines/basic/gen3c/cache_3d.py`
+- `fastvideo/pipelines/basic/gen3c/depth_estimation.py`
+- `fastvideo/models/vaes/gen3c_tokenizer_vae.py`
+
+## References
+
+- [GEN3C Paper](https://arxiv.org/abs/2503.03751)
+- [Official Repository](https://github.com/nv-tlabs/GEN3C)
+- [Official Checkpoint (raw)](https://huggingface.co/nvidia/GEN3C-Cosmos-7B)
@@ -73,6 +73,7 @@ pipeline initialization and sampling.
 | Matrix Game 2.0 Base | `FastVideo/Matrix-Game-2.0-Base-Diffusers` | 352x640 | ⭕ | ⭕ | ⭕ | ⭕ | ⭕ |
 | Matrix Game 2.0 GTA | `FastVideo/Matrix-Game-2.0-GTA-Diffusers` | 352x640 | ⭕ | ⭕ | ⭕ | ⭕ | ⭕ |
 | Matrix Game 2.0 TempleRun | `FastVideo/Matrix-Game-2.0-TempleRun-Diffusers` | 352x640 | ⭕ | ⭕ | ⭕ | ⭕ | ⭕ |
+| GEN3C Cosmos 7B | `FastVideo/GEN3C-Cosmos-7B-Diffusers` | 704px1280p | ❌ | ❌ | ❌ | ⭕ | ⭕ |

 **Note**: Wan2.2 TI2V 5B has some quality issues when performing I2V generation. We are working on fixing this issue.

@@ -85,6 +86,11 @@ The authoritative source for model-ID recognition is
 `fastvideo/registry.py`. If a model ID is registered there, FastVideo can
 resolve default pipeline and sampling configuration for it.

+**Note (GEN3C)**: The official `nvidia/GEN3C-Cosmos-7B` repo provides a raw
+`model.pt` checkpoint. Use a Diffusers-format repo (for example,
+`FastVideo/GEN3C-Cosmos-7B-Diffusers`) or convert locally with
+`scripts/checkpoint_conversion/convert_gen3c_to_fastvideo.py`.
+
 ## Special requirements

 ### Sliding Tile Attention
@@ -28,6 +28,11 @@ For an example running DMD+VSA inference:
 python examples/inference/basic/basic_dmd.py
 ```

+For the typed config/request path added during the inference API refactor:
+```
+python examples/inference/basic/basic_dmd_new_api.py
+```
+
 ## Basic Walkthrough

 All you need to generate videos using multi-gpus from state-of-the-art diffusion pipelines is the following few lines!
@@ -0,0 +1,98 @@
+import os
+import time
+
+from fastvideo import VideoGenerator
+from fastvideo.api import (
+    EngineConfig,
+    GenerationRequest,
+    GeneratorConfig,
+    OffloadConfig,
+    OutputConfig,
+    PipelineSelection,
+)
+
+OUTPUT_PATH = "video_samples_dmd2_typed"
+
+
+def main():
+    os.environ["FASTVIDEO_ATTENTION_BACKEND"] = "VIDEO_SPARSE_ATTN"
+
+    model_name = "FastVideo/FastWan2.1-T2V-1.3B-Diffusers"
+    generator_config = GeneratorConfig(
+        model_path=model_name,
+        engine=EngineConfig(
+            num_gpus=1,
+            use_fsdp_inference=False,
+            offload=OffloadConfig(
+                text_encoder=True,
+                pin_cpu_memory=True,
+                dit=False,
+                vae=False,
+            ),
+        ),
+        # PR 2 still routes a few advanced inference knobs through the
+        # compatibility bridge until they get first-class typed fields.
+        pipeline=PipelineSelection(
+            experimental={
+                "VSA_sparsity": 0.8,
+            },
+        ),
+    )
+
+    load_start_time = time.perf_counter()
+    generator = VideoGenerator.from_config(generator_config)
+    load_end_time = time.perf_counter()
+    load_time = load_end_time - load_start_time
+
+    prompt = (
+        "A neon-lit alley in futuristic Tokyo during a heavy rainstorm at night. "
+        "The puddles reflect glowing signs in kanji, advertising ramen, karaoke, "
+        "and VR arcades. A woman in a translucent raincoat walks briskly with an "
+        "LED umbrella. Steam rises from a street food cart, and a cat darts "
+        "across the screen. Raindrops are visible on the camera lens, creating "
+        "a cinematic bokeh effect."
+    )
+    request = GenerationRequest(
+        prompt=prompt,
+        output=OutputConfig(
+            output_path=OUTPUT_PATH,
+            save_video=True,
+            return_frames=False,
+        ),
+    )
+
+    start_time = time.perf_counter()
+    result = generator.generate(request)
+    end_time = time.perf_counter()
+    gen_time = end_time - start_time
+
+    prompt2 = (
+        "A majestic lion strides across the golden savanna, its powerful frame "
+        "glistening under the warm afternoon sun. The tall grass ripples gently "
+        "in the breeze, enhancing the lion's commanding presence. The tone is "
+        "vibrant, embodying the raw energy of the wild. Low angle, steady "
+        "tracking shot, cinematic."
+    )
+    request2 = GenerationRequest(
+        prompt=prompt2,
+        output=OutputConfig(
+            output_path=OUTPUT_PATH,
+            save_video=True,
+            return_frames=False,
+        ),
+    )
+
+    start_time = time.perf_counter()
+    result2 = generator.generate(request2)
+    end_time = time.perf_counter()
+    gen_time2 = end_time - start_time
+
+    print(f"Time taken to load model: {load_time} seconds")
+    print(f"Time taken to generate video: {gen_time} seconds")
+    print(f"First output written to: {result.video_path}")
+    print(f"Time taken to generate video2: {gen_time2} seconds")
+    print(f"Second output written to: {result2.video_path}")
+
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,109 @@
+"""
+GEN3C: 3D-aware camera-controlled video generation.
+
+This example generates a video from a single input image with camera control.
+The pipeline uses MoGe depth estimation, 3D point cloud forward warping,
+and the GEN3C diffusion model.
+
+Requirements:
+  1. Install MoGe:
+     pip install git+https://github.com/microsoft/MoGe.git
+     If you hit `ImportError: libGL.so.1`, install:
+     sudo apt-get update && sudo apt-get install -y libgl1 libglib2.0-0 libsm6 libxext6 libxrender1
+  2. Download and convert weights:
+     huggingface-cli download nvidia/GEN3C-Cosmos-7B --local-dir official_weights/GEN3C-Cosmos-7B
+     python scripts/checkpoint_conversion/convert_gen3c_to_fastvideo.py \
+       --source ./official_weights/GEN3C-Cosmos-7B/model.pt \
+       --output ./converted_weights/GEN3C-Cosmos-7B \
+       --components-source nvidia/Cosmos-Predict2-2B-Video2World
+  3. Provide an input image for 3D-conditioned generation.
+"""
+
+import argparse
+
+from fastvideo import VideoGenerator
+
+
+def main():
+    parser = argparse.ArgumentParser(description="GEN3C video generation")
+    parser.add_argument("--model_path",
+                        type=str,
+                        default="converted_weights/GEN3C-Cosmos-7B")
+    parser.add_argument("--image_path",
+                        type=str,
+                        default=None,
+                        help="Input image for 3D cache conditioning")
+    parser.add_argument("--prompt",
+                        type=str,
+                        default="A slow camera pan over a sunlit landscape.")
+    parser.add_argument(
+        "--negative_prompt",
+        type=str,
+        default=(
+            "The video captures a series of frames showing ugly scenes, static with no motion, motion blur, "
+            "over-saturation, shaky footage, low resolution, grainy texture, pixelated images, poorly lit areas, "
+            "underexposed and overexposed scenes, poor color balance, washed out colors, choppy sequences, "
+            "jerky movements, low frame rate, artifacting, color banding, unnatural transitions, outdated special "
+            "effects, fake elements, unconvincing visuals, poorly edited content, jump cuts, visual noise, and "
+            "flickering. Overall, the video is of poor quality."
+        ),
+    )
+    parser.add_argument("--trajectory",
+                        type=str,
+                        default="left",
+                        choices=[
+                            "left", "right", "up", "down", "zoom_in",
+                            "zoom_out", "clockwise", "counterclockwise", "none"
+                        ])
+    parser.add_argument("--movement_distance", type=float, default=0.3)
+    parser.add_argument("--camera_rotation",
+                        type=str,
+                        default="center_facing",
+                        choices=[
+                            "center_facing", "no_rotation",
+                            "trajectory_aligned"
+                        ])
+    parser.add_argument("--height", type=int, default=704)
+    parser.add_argument("--width", type=int, default=1280)
+    parser.add_argument("--num_frames", type=int, default=121)
+    parser.add_argument("--num_inference_steps", type=int, default=35)
+    parser.add_argument("--guidance_scale", type=float, default=1.0)
+    parser.add_argument("--output_path",
+                        type=str,
+                        default="outputs_video/gen3c.mp4")
+    parser.add_argument("--seed", type=int, default=42)
+    args = parser.parse_args()
+
+    generator = VideoGenerator.from_pretrained(
+        args.model_path,
+        num_gpus=1,
+        use_fsdp_inference=False,
+        dit_cpu_offload=False,
+        vae_cpu_offload=True,
+        text_encoder_cpu_offload=True,
+        pin_cpu_memory=True,
+    )
+
+    video = generator.generate_video(
+        args.prompt,
+        negative_prompt=args.negative_prompt,
+        image_path=args.image_path,
+        trajectory_type=args.trajectory,
+        movement_distance=args.movement_distance,
+        camera_rotation=args.camera_rotation,
+        height=args.height,
+        width=args.width,
+        num_frames=args.num_frames,
+        num_inference_steps=args.num_inference_steps,
+        guidance_scale=args.guidance_scale,
+        fps=24,
+        seed=args.seed,
+        output_path=args.output_path,
+        save_video=True,
+    )
+
+    generator.shutdown()
+
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,93 @@
+# 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.
@@ -0,0 +1,21 @@
+# 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/`.
@@ -0,0 +1,72 @@
+# 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
@@ -0,0 +1,68 @@
+# 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
@@ -0,0 +1,89 @@
+# Knowledge Distillation: Wan 2.1 Causal T2V 1.3B
+#
+# Trains a causal student (1.3B) with per-frame block-quantized MSE loss
+# on teacher (14B) ODE trajectories. The resulting checkpoint serves as
+# the ode_init weight initialization for downstream Self-Forcing training.
+#
+# The same teacher_path_cache can be shared with kd_wan2.1_t2v_1.3B.yaml
+# since the cache stores teacher trajectories independent of the student.
+#
+# After training, export the ode_init checkpoint:
+#   python -m fastvideo.train.entrypoint.dcp_to_diffusers \
+#       --role student \
+#       --output_dir outputs/wan2.1_causal_kd_ode_init/diffusers \
+#       --checkpoint_dir outputs/wan2.1_causal_kd_ode_init/latest
+#
+# The exported model.safetensors can then be passed as:
+#   transformer_override_safetensor in a Self-Forcing config.
+#
+# Remove the 'teacher' block once the cache is complete to avoid
+# loading the 14B model on subsequent runs.
+
+models:
+  student:
+    _target_: fastvideo.train.models.wan.WanCausalModel
+    init_from: wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers
+    trainable: true
+  teacher:                          # remove once cache is complete
+    _target_: fastvideo.train.models.wan.WanModel
+    init_from: Wan-AI/Wan2.1-T2V-14B-Diffusers
+    trainable: false
+    disable_custom_init_weights: true
+
+method:
+  _target_: fastvideo.train.methods.knowledge_distillation.kd.KDCausalMethod
+  teacher_path_cache: data/kd_cache/wan14b_4step
+  t_list: [999, 937, 833, 624, 0]   # student inference schedule (num_steps+1)
+  student_sample_steps: 4
+  teacher_inference_steps: 48       # teacher ODE steps; nearest indices selected for t_list
+  teacher_guidance_scale: 3.5       # CFG scale during teacher path generation
+  num_frames_per_block: 3           # frames sharing the same noise level (per-frame KD)
+
+training:
+  distributed:
+    num_gpus: 8
+    sp_size: 1
+    tp_size: 1
+    hsdp_replicate_dim: 1
+    hsdp_shard_dim: 8
+
+  data:
+    data_path: data/Wan-Syn_77x448x832_600k
+    dataloader_num_workers: 4
+    train_batch_size: 2
+    training_cfg_rate: 0.0
+    seed: 1000
+    num_latent_t: 20
+    num_height: 448
+    num_width: 832
+    num_frames: 77
+
+  optimizer:
+    learning_rate: 7e-5
+    betas: [0.9, 0.999]
+    weight_decay: 0.0
+    lr_scheduler: constant
+    lr_warmup_steps: 0
+
+  loop:
+    max_train_steps: 10000
+    gradient_accumulation_steps: 1
+
+  checkpoint:
+    output_dir: outputs/wan2.1_causal_kd_ode_init
+    training_state_checkpointing_steps: 1000
+    checkpoints_total_limit: 3
+
+  tracker:
+    project_name: kd-wan
+    run_name: wan2.1_causal_kd_ode_init
+
+  model:
+    enable_gradient_checkpointing_type: full
+
+callbacks:
+  grad_clip:
+    max_grad_norm: 1.0
+
+pipeline:
+  flow_shift: 5
@@ -0,0 +1,85 @@
+# Knowledge Distillation: Wan 2.1 T2V 1.3B
+#
+# Trains student (1.3B) with MSE loss on teacher (14B) ODE trajectories.
+# The resulting checkpoint serves as the ode_init weight initialization
+# for downstream Self-Forcing training.
+#
+# After training, export the ode_init checkpoint:
+#   python -m fastvideo.train.entrypoint.dcp_to_diffusers \
+#       --role student \
+#       --output_dir outputs/wan2.1_kd_ode_init/diffusers \
+#       --checkpoint_dir outputs/wan2.1_kd_ode_init/latest
+#
+# The exported model.safetensors can then be passed as:
+#   transformer_override_safetensor in a Self-Forcing config.
+#
+# Remove the 'teacher' block once the cache is complete to avoid
+# loading the 14B model on subsequent runs.
+
+models:
+  student:
+    _target_: fastvideo.train.models.wan.WanModel
+    init_from: Wan-AI/Wan2.1-T2V-1.3B-Diffusers
+    trainable: true
+  teacher:                          # remove once cache is complete
+    _target_: fastvideo.train.models.wan.WanModel
+    init_from: Wan-AI/Wan2.1-T2V-14B-Diffusers
+    trainable: false
+    disable_custom_init_weights: true
+
+method:
+  _target_: fastvideo.train.methods.knowledge_distillation.kd.KDMethod
+  teacher_path_cache: data/kd_cache/wan14b_4step
+  t_list: [999, 937, 833, 624, 0]   # student inference schedule (num_steps+1)
+  student_sample_steps: 4
+  teacher_inference_steps: 48       # teacher ODE steps; nearest indices selected for t_list
+  teacher_guidance_scale: 3.5       # CFG scale during teacher path generation
+
+training:
+  distributed:
+    num_gpus: 8
+    sp_size: 1
+    tp_size: 1
+    hsdp_replicate_dim: 1
+    hsdp_shard_dim: 8
+
+  data:
+    data_path: data/Wan-Syn_77x448x832_600k
+    dataloader_num_workers: 4
+    train_batch_size: 2
+    training_cfg_rate: 0.0
+    seed: 1000
+    num_latent_t: 20
+    num_height: 448
+    num_width: 832
+    num_frames: 77
+
+  optimizer:
+    learning_rate: 7e-5
+    betas: [0.9, 0.999]
+    weight_decay: 0.0
+    lr_scheduler: constant
+    lr_warmup_steps: 0
+
+  loop:
+    max_train_steps: 10000
+    gradient_accumulation_steps: 1
+
+  checkpoint:
+    output_dir: outputs/wan2.1_kd_ode_init
+    training_state_checkpointing_steps: 1000
+    checkpoints_total_limit: 3
+
+  tracker:
+    project_name: kd-wan
+    run_name: wan2.1_kd_ode_init
+
+  model:
+    enable_gradient_checkpointing_type: full
+
+callbacks:
+  grad_clip:
+    max_grad_norm: 1.0
+
+pipeline:
+  flow_shift: 5
@@ -0,0 +1,78 @@
+# 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
@@ -5,12 +5,12 @@
 #   bash examples/train/run.sh <config.yaml> [--dotted.key value ...]
 #
 # Examples:
-#   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 \
+#   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 \
 #       --training.distributed.num_gpus 4 \
 #       --training.optimizer.learning_rate 1e-5
-#   bash examples/train/run.sh examples/train/distill_wan2.1_t2v_1.3B_dmd2.yaml \
+#   bash examples/train/run.sh examples/train/configs/distribution_matching/wan/dmd2_t2v.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).
@@ -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/distill_wan2.1_t2v_1.3B_dmd2.yaml 4 \
+#   bash examples/train/run_slurm.sh examples/train/configs/distribution_matching/wan/dmd2_t2v.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
@@ -0,0 +1,13 @@
+# 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/`.
@@ -0,0 +1,92 @@
+"""Decode 'real' latents from a KD cache to videos.
+
+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 \
+        --cache_dir data/kd_test_cache_small \
+        --out_dir   data/kd_test_cache_small/video \
+        --fps 16
+"""
+
+import argparse
+from pathlib import Path
+
+import imageio
+import numpy as np
+import torch
+from diffusers import AutoencoderKLWan
+
+MODEL_ID = "Wan-AI/Wan2.1-T2V-1.3B-Diffusers"
+
+
+def denormalize(latents: torch.Tensor, vae: AutoencoderKLWan) -> torch.Tensor:
+    """Inverse of normalize_dit_input('wan', ...).
+
+    normalize_dit_input: normalized = (latent - mean) * (1/std)
+    inverse:             latent     = normalized * std + mean
+    """
+    cfg = vae.config
+    mean = torch.tensor(cfg.latents_mean, dtype=latents.dtype,
+                        device=latents.device).view(1, -1, 1, 1, 1)
+    std = torch.tensor(cfg.latents_std, dtype=latents.dtype,
+                       device=latents.device).view(1, -1, 1, 1, 1)
+    return latents * std + mean
+
+
+def decode_real(pt_path: Path, vae: AutoencoderKLWan,
+                device: torch.device) -> np.ndarray:
+    """Load one .pt cache file and decode its 'real' latent to uint8 [T,H,W,3]."""
+    d = torch.load(pt_path, weights_only=True, map_location="cpu")
+    real = d["real"].float()  # [T, C, H, W]  (normalized)
+
+    # [T, C, H, W] → [1, C, T, H, W]
+    latent = real.permute(1, 0, 2, 3).unsqueeze(0).to(device)
+    latent = denormalize(latent, vae)
+
+    with torch.no_grad():
+        video = vae.decode(latent).sample  # [1, 3, T_px, H_px, W_px]
+
+    video = video.squeeze(0).permute(1, 2, 3, 0)  # [T, H, W, 3]
+    video = ((video.clamp(-1, 1) + 1) / 2 * 255).cpu().float().numpy()
+    return video.astype(np.uint8)
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--cache_dir", default="data/kd_test_cache_small")
+    parser.add_argument("--out_dir", default="data/kd_test_cache_small/video")
+    parser.add_argument("--model_id", default=MODEL_ID)
+    parser.add_argument("--fps", type=int, default=16)
+    parser.add_argument("--overwrite", action="store_true")
+    args = parser.parse_args()
+
+    samples_dir = Path(args.cache_dir) / "samples"
+    out_dir = Path(args.out_dir)
+    out_dir.mkdir(parents=True, exist_ok=True)
+
+    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+    print(f"Loading VAE from {args.model_id} ...")
+    vae = AutoencoderKLWan.from_pretrained(
+        args.model_id, subfolder="vae", torch_dtype=torch.float32)
+    vae.eval().to(device)
+
+    pts = sorted(samples_dir.glob("*.pt"))
+    print(f"Found {len(pts)} samples in {samples_dir}")
+
+    for pt in pts:
+        out_path = out_dir / (pt.stem + ".mp4")
+        if out_path.exists() and not args.overwrite:
+            print(f"  skip {pt.name} (exists)")
+            continue
+        video_np = decode_real(pt, vae, device)
+        imageio.mimsave(str(out_path), video_np, fps=args.fps)
+        print(f"  {pt.name} → {out_path.name}  {video_np.shape}")
+
+    print(f"Done. Videos in {out_dir}/")
+
+
+if __name__ == "__main__":
+    main()
@@ -0,0 +1,91 @@
+# KD Test: Causal Wan 1.3B ODE-init
+#
+# Smoke-test for KDCausalMethod on a tiny 16-sample dataset.
+# Uses 1.3B as teacher (instead of 14B) so everything fits on 1 GPU.
+#
+# 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
+#
+# Step 2 — Export ode_init checkpoint:
+#   python -m fastvideo.train.entrypoint.dcp_to_diffusers \
+#       --role student \
+#       --checkpoint_dir outputs/kdtest/kd_causal/latest \
+#       --output_dir outputs/kdtest/kd_ode_init
+#
+# 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
+#
+# Success criteria:
+#   - Cache generates 16 .pt files under data/kd_test_cache/samples/
+#   - kd_loss decreases steadily and approaches ~0 by step 200
+
+models:
+  student:
+    _target_: fastvideo.train.models.wan.WanCausalModel
+    init_from: wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers
+    trainable: true
+  teacher:                          # using 1.3B for single-GPU test; swap to 14B for real runs
+    _target_: fastvideo.train.models.wan.WanModel
+    init_from: Wan-AI/Wan2.1-T2V-1.3B-Diffusers
+    trainable: false
+    disable_custom_init_weights: true
+
+method:
+  _target_: fastvideo.train.methods.knowledge_distillation.kd.KDCausalMethod
+  teacher_path_cache: data/kd_test_cache_small
+  t_list: [995, 937, 833, 625, 0]
+  student_sample_steps: 4
+  teacher_inference_steps: 48
+  teacher_guidance_scale: 3.5
+  num_frames_per_block: 3
+
+training:
+  distributed:
+    num_gpus: 8
+    sp_size: 1
+    tp_size: 1
+    hsdp_replicate_dim: 1
+    hsdp_shard_dim: 8
+
+  data:
+    data_path: data/Wan-Syn_small
+    dataloader_num_workers: 0
+    train_batch_size: 1
+    training_cfg_rate: 0.0
+    seed: 42
+    num_latent_t: 20
+    num_height: 448
+    num_width: 832
+    num_frames: 77
+
+  optimizer:
+    learning_rate: 7e-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/kdtest/kd_causal
+    training_state_checkpointing_steps: 300
+    checkpoints_total_limit: 1
+
+  tracker:
+    project_name: kd-test
+    run_name: kd_causal_smoke
+
+  model:
+    enable_gradient_checkpointing_type: full
+
+callbacks:
+  grad_clip:
+    max_grad_norm: 1.0
+
+pipeline:
+  flow_shift: 5
@@ -0,0 +1,21 @@
+#!/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."
@@ -0,0 +1,111 @@
+# 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.
+#
+# Prerequisite — export ode_init from the KD run:
+#   python -m fastvideo.train.entrypoint.dcp_to_diffusers \
+#       --role student \
+#       --checkpoint_dir outputs/kdtest/kd_causal/latest \
+#       --output_dir outputs/kdtest/kd_ode_init
+#
+# 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
+#
+# Success criteria:
+#   - generator_loss and critic_loss both finite at step 0
+#   - Validation videos logged to tracker every 20 steps
+#   - No crash through max_train_steps=100
+
+models:
+  student:
+    _target_: fastvideo.train.models.wan.WanCausalModel
+    init_from: wlsaidhi/SFWan2.1-T2V-1.3B-Diffusers
+    trainable: true
+    transformer_override_safetensor: outputs/kdtest/kd_ode_init/transformer/model.safetensors
+  teacher:
+    _target_: fastvideo.train.models.wan.WanModel
+    init_from: Wan-AI/Wan2.1-T2V-14B-Diffusers
+    trainable: false
+    disable_custom_init_weights: true
+  critic:
+    _target_: fastvideo.train.models.wan.WanModel
+    init_from: Wan-AI/Wan2.1-T2V-1.3B-Diffusers
+    trainable: true
+    disable_custom_init_weights: true
+
+method:
+  _target_: fastvideo.train.methods.distribution_matching.self_forcing.SelfForcingMethod
+  rollout_mode: simulate
+  generator_update_interval: 5
+  real_score_guidance_scale: 4.5
+  dmd_denoising_steps: [1000, 750, 500, 250]
+
+  chunk_size: 3
+  student_sample_type: sde
+  context_noise: 0.0
+  enable_gradient_in_rollout: true
+  start_gradient_frame: 0
+
+  # Critic optimizer
+  fake_score_learning_rate: 8.0e-6
+  fake_score_betas: [0.0, 0.999]
+  fake_score_lr_scheduler: constant
+
+training:
+  distributed:
+    num_gpus: 8
+    sp_size: 1
+    tp_size: 1
+    hsdp_replicate_dim: 1
+    hsdp_shard_dim: 8
+
+  data:
+    data_path: data/Wan-Syn_small
+    dataloader_num_workers: 0
+    train_batch_size: 1
+    training_cfg_rate: 0.0
+    seed: 42
+    num_latent_t: 20
+    num_height: 448
+    num_width: 832
+    num_frames: 77
+
+  optimizer:
+    learning_rate: 1e-5
+    betas: [0.0, 0.999]
+    weight_decay: 0.01
+    lr_scheduler: constant
+    lr_warmup_steps: 0
+
+  loop:
+    max_train_steps: 100
+    gradient_accumulation_steps: 1
+
+  checkpoint:
+    output_dir: outputs/kdtest/self_forcing_causal
+    training_state_checkpointing_steps: 100
+    checkpoints_total_limit: 1
+
+  tracker:
+    project_name: kd-test
+    run_name: self_forcing_causal_smoke
+
+  model:
+    enable_gradient_checkpointing_type: full
+
+callbacks:
+  grad_clip:
+    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
+    every_steps: 20
+    sampling_steps: [4]
+    sampling_timesteps: [1000, 750, 500, 250]
+    num_frames: 69
+    guidance_scale: 6.0
+
+pipeline:
+  flow_shift: 5
@@ -0,0 +1,19 @@
+# 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
+```
@@ -0,0 +1,20 @@
+{
+  "data": [
+    {
+      "caption": "A stylish woman walks down a Tokyo street filled with warm glowing neon and animated city signage. She wears a black leather jacket, a long red dress, and black boots, and carries a black purse. She walks confidently and casually. The street is damp and reflective, creating a mirror effect of the colorful lights.",
+      "video_path": null,
+      "num_inference_steps": 4,
+      "height": 448,
+      "width": 832,
+      "num_frames": 69
+    },
+    {
+      "caption": "A white and orange tabby cat is seen happily darting through a dense garden, as if chasing something. Its eyes are wide and happy as it jogs forward, scanning the branches, flowers, and leaves as it walks. The scene is captured from a ground-level angle, giving a low and intimate perspective.",
+      "video_path": null,
+      "num_inference_steps": 4,
+      "height": 448,
+      "width": 832,
+      "num_frames": 69
+    }
+  ]
+}
@@ -10,6 +10,35 @@ set -ex

 echo "Building fastvideo-kernel..."

+# ---------------------------------------------------------------------------
+# Neutralise conda-injected compiler toolchains.
+#
+# Conda compiler packages (gcc_linux-aarch64, gxx_linux-64, etc.) set
+# CMAKE_ARGS, CFLAGS, CXXFLAGS, and LDFLAGS on activation.  When multiple
+# toolchains are installed the variables can reference a *cross*-compiler
+# that doesn't match the host (e.g. aarch64-conda-linux-gnu-c++ on x86_64).
+# Even when the correct toolchain is active, the flags it injects
+# (-march=nocona, -mtune=haswell, …) can conflict with nvcc's host-compiler
+# expectations.  Clear them so CMake discovers the system compiler instead.
+# ---------------------------------------------------------------------------
+if [[ -n "${CONDA_PREFIX:-}" ]]; then
+    _need_clean=0
+    # Detect conda cross-compiler that doesn't match the host.
+    _host_arch="$(uname -m)"
+    if [[ "${CXX:-}" == *"conda"* ]] || [[ "${CC:-}" == *"conda"* ]]; then
+        _need_clean=1
+    fi
+    if [[ "${CMAKE_ARGS:-}" == *"conda"* ]]; then
+        _need_clean=1
+    fi
+    if (( _need_clean )); then
+        echo "NOTE: Clearing conda-injected compiler settings (CC/CXX/CMAKE_ARGS/CFLAGS/...)"
+        echo "      to use the system compiler for CUDA extension builds."
+        unset CC CXX CMAKE_ARGS CFLAGS CXXFLAGS LDFLAGS
+    fi
+    unset _need_clean _host_arch
+fi
+
 # Ensure submodules are initialized if needed (tk)
 git submodule update --init --recursive

@@ -23,7 +23,7 @@ classifiers = [
 ]
 dependencies = [
    "torch>=2.5.0",
-    "triton>=2.0.0",
+    "triton>=2.0.0; sys_platform == 'linux'",
 ]

 [project.urls]
@@ -0,0 +1,65 @@
+# SPDX-License-Identifier: Apache-2.0
+from fastvideo.api.schema import (
+    CompileConfig,
+    ComponentConfig,
+    ContinuationState,
+    EngineConfig,
+    GenerationPlan,
+    GenerationRequest,
+    GeneratorConfig,
+    InputConfig,
+    OffloadConfig,
+    OutputConfig,
+    ParallelismConfig,
+    PipelineSelection,
+    PlannedStage,
+    QuantizationConfig,
+    RequestRuntimeConfig,
+    RunConfig,
+    SamplingConfig,
+    ServeConfig,
+    ServerConfig,
+)
+from fastvideo.api.errors import ConfigValidationError
+from fastvideo.api.overrides import apply_overrides, parse_cli_overrides
+from fastvideo.api.parser import (
+    config_to_dict,
+    load_config,
+    load_raw_config,
+    load_run_config,
+    load_serve_config,
+    parse_config,
+)
+from fastvideo.api.results import GenerationResult
+
+__all__ = [
+    "CompileConfig",
+    "ComponentConfig",
+    "ContinuationState",
+    "ConfigValidationError",
+    "EngineConfig",
+    "GenerationResult",
+    "GenerationPlan",
+    "GenerationRequest",
+    "GeneratorConfig",
+    "InputConfig",
+    "OffloadConfig",
+    "OutputConfig",
+    "ParallelismConfig",
+    "PipelineSelection",
+    "PlannedStage",
+    "QuantizationConfig",
+    "RequestRuntimeConfig",
+    "RunConfig",
+    "SamplingConfig",
+    "ServeConfig",
+    "ServerConfig",
+    "apply_overrides",
+    "config_to_dict",
+    "load_config",
+    "load_raw_config",
+    "load_run_config",
+    "load_serve_config",
+    "parse_cli_overrides",
+    "parse_config",
+]
@@ -0,0 +1,507 @@
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+from collections.abc import Mapping
+from copy import deepcopy
+from dataclasses import fields, is_dataclass
+from pathlib import Path
+from typing import Any
+
+from fastvideo.api.overrides import apply_overrides, parse_cli_overrides
+from fastvideo.api.parser import config_to_dict, load_raw_config, parse_config
+from fastvideo.api.request_metadata import (
+    EXPLICIT_REQUEST_ATTR,
+    bind_generation_request_raw,
+    refresh_generation_request_raw,
+)
+from fastvideo.api.schema import (
+    GenerationRequest,
+    GeneratorConfig,
+    InputConfig,
+    OutputConfig,
+    RequestRuntimeConfig,
+    SamplingConfig,
+)
+from fastvideo.configs.sample import SamplingParam
+from fastvideo.fastvideo_args import FastVideoArgs
+from fastvideo.utils import shallow_asdict
+
+_INPUT_FIELD_NAMES = {field.name for field in fields(InputConfig)}
+_SAMPLING_FIELD_NAMES = {field.name for field in fields(SamplingConfig)}
+_RUNTIME_FIELD_NAMES = {field.name for field in fields(RequestRuntimeConfig)}
+_OUTPUT_FIELD_NAMES = {field.name for field in fields(OutputConfig)}
+_MISSING = object()
+_LEGACY_REQUEST_ALIASES = {
+    "neg_prompt": "negative_prompt",
+}
+_REQUEST_PIPELINE_OVERRIDE_FIELDS = frozenset({
+    "embedded_cfg_scale",
+})
+
+
+def normalize_generator_config(config: GeneratorConfig | Mapping[str, Any], ) -> GeneratorConfig:
+    if isinstance(config, GeneratorConfig):
+        return config
+    return parse_config(GeneratorConfig, config)
+
+
+def load_generator_config_from_file(
+    path: str | Path,
+    overrides: list[str] | Mapping[str, Any] | None = None,
+) -> GeneratorConfig:
+    raw = load_raw_config(path)
+    normalized_overrides = _normalize_overrides(overrides)
+
+    if _looks_like_run_or_serve_config(raw):
+        if normalized_overrides:
+            raw = apply_overrides(raw, normalized_overrides)
+        return parse_config(GeneratorConfig, raw["generator"])
+
+    if normalized_overrides:
+        adjusted = normalized_overrides
+        if all(key.startswith("generator.") for key in adjusted):
+            adjusted = {key[len("generator."):]: value for key, value in adjusted.items()}
+        raw = apply_overrides(raw, adjusted)
+
+    return parse_config(GeneratorConfig, raw)
+
+
+def legacy_from_pretrained_to_config(
+    model_path: str,
+    kwargs: Mapping[str, Any],
+) -> GeneratorConfig:
+    raw: dict[str, Any] = {"model_path": model_path}
+    engine: dict[str, Any] = {}
+    parallelism: dict[str, Any] = {}
+    offload: dict[str, Any] = {}
+    compile_config: dict[str, Any] = {}
+    pipeline: dict[str, Any] = {}
+    components: dict[str, Any] = {}
+    quantization: dict[str, Any] = {}
+    experimental: dict[str, Any] = {}
+
+    for key, value in kwargs.items():
+        if key == "revision":
+            raw["revision"] = value
+        elif key == "trust_remote_code":
+            raw["trust_remote_code"] = value
+        elif key == "num_gpus":
+            engine["num_gpus"] = value
+        elif key == "distributed_executor_backend":
+            engine["execution_backend"] = value
+        elif key in {"tp_size", "sp_size", "hsdp_replicate_dim", "hsdp_shard_dim", "dist_timeout"}:
+            parallelism[key] = value
+        elif key == "dit_cpu_offload":
+            offload["dit"] = value
+        elif key == "dit_layerwise_offload":
+            offload["dit_layerwise"] = value
+        elif key == "text_encoder_cpu_offload":
+            offload["text_encoder"] = value
+        elif key == "image_encoder_cpu_offload":
+            offload["image_encoder"] = value
+        elif key == "vae_cpu_offload":
+            offload["vae"] = value
+        elif key == "pin_cpu_memory":
+            offload["pin_cpu_memory"] = value
+        elif key == "enable_torch_compile":
+            compile_config["enabled"] = value
+        elif key == "torch_compile_kwargs":
+            compile_config["kwargs"] = deepcopy(value)
+        elif key in {"enable_stage_verification", "use_fsdp_inference", "disable_autocast"}:
+            engine[key] = value
+        elif key == "override_text_encoder_quant":
+            quantization["text_encoder_quant"] = value
+        elif key == "workload_type":
+            pipeline["workload_type"] = value
+        elif key == "lora_path":
+            components["lora_path"] = value
+        elif key == "override_pipeline_cls_name":
+            components["override_pipeline_cls_name"] = value
+        elif key == "override_transformer_cls_name":
+            components["override_transformer_cls_name"] = value
+        elif key == "pipeline_config":
+            if isinstance(value, str):
+                components["pipeline_config_path"] = value
+            else:
+                experimental[key] = deepcopy(value)
+        elif key == "override_text_encoder_safetensors":
+            components["text_encoder_weights"] = value
+        elif key == "init_weights_from_safetensors":
+            components["transformer_weights"] = value
+        elif key == "init_weights_from_safetensors_2":
+            components["transformer_2_weights"] = value
+        else:
+            experimental[key] = deepcopy(value)
+
+    if parallelism:
+        engine["parallelism"] = parallelism
+    if offload:
+        engine["offload"] = offload
+    if compile_config:
+        engine["compile"] = compile_config
+    if quantization:
+        engine["quantization"] = quantization
+    if engine:
+        raw["engine"] = engine
+
+    if components:
+        pipeline["components"] = components
+    if experimental:
+        pipeline["experimental"] = experimental
+    if pipeline:
+        raw["pipeline"] = pipeline
+
+    return parse_config(GeneratorConfig, raw)
+
+
+def generator_config_to_fastvideo_args(config: GeneratorConfig | Mapping[str, Any], ) -> FastVideoArgs:
+    normalized = normalize_generator_config(config)
+    unsupported = []
+    if normalized.pipeline.profile is not None:
+        unsupported.append("pipeline.profile")
+    if normalized.pipeline.profile_version is not None:
+        unsupported.append("pipeline.profile_version")
+    if normalized.pipeline.components.config_root is not None:
+        unsupported.append("pipeline.components.config_root")
+    if normalized.pipeline.components.vae_weights is not None:
+        unsupported.append("pipeline.components.vae_weights")
+    if normalized.pipeline.components.upsampler_weights is not None:
+        unsupported.append("pipeline.components.upsampler_weights")
+    if unsupported:
+        joined = ", ".join(unsupported)
+        raise NotImplementedError(f"VideoGenerator compatibility adapter does not support {joined} yet")
+
+    engine = normalized.engine
+    kwargs: dict[str, Any] = {
+        "model_path": normalized.model_path,
+        "revision": normalized.revision,
+        "trust_remote_code": normalized.trust_remote_code,
+        "num_gpus": engine.num_gpus,
+        "distributed_executor_backend": engine.execution_backend,
+        "tp_size": engine.parallelism.tp_size,
+        "sp_size": engine.parallelism.sp_size,
+        "hsdp_replicate_dim": engine.parallelism.hsdp_replicate_dim,
+        "hsdp_shard_dim": engine.parallelism.hsdp_shard_dim,
+        "dist_timeout": engine.parallelism.dist_timeout,
+        "dit_cpu_offload": engine.offload.dit,
+        "dit_layerwise_offload": engine.offload.dit_layerwise,
+        "text_encoder_cpu_offload": engine.offload.text_encoder,
+        "image_encoder_cpu_offload": engine.offload.image_encoder,
+        "vae_cpu_offload": engine.offload.vae,
+        "pin_cpu_memory": engine.offload.pin_cpu_memory,
+        "enable_torch_compile": engine.compile.enabled,
+        "torch_compile_kwargs": deepcopy(engine.compile.kwargs),
+        "enable_stage_verification": engine.enable_stage_verification,
+        "use_fsdp_inference": engine.use_fsdp_inference,
+        "disable_autocast": engine.disable_autocast,
+    }
+    if normalized.pipeline.workload_type is not None:
+        kwargs["workload_type"] = normalized.pipeline.workload_type
+
+    quantization = engine.quantization
+    if quantization is not None and quantization.text_encoder_quant is not None:
+        kwargs["override_text_encoder_quant"] = quantization.text_encoder_quant
+    if quantization is not None and quantization.transformer_quant is not None:
+        kwargs["transformer_quant"] = quantization.transformer_quant
+
+    components = normalized.pipeline.components
+    if components.pipeline_config_path is not None:
+        kwargs["pipeline_config"] = components.pipeline_config_path
+    if components.lora_path is not None:
+        kwargs["lora_path"] = components.lora_path
+    if components.override_pipeline_cls_name is not None:
+        kwargs["override_pipeline_cls_name"] = components.override_pipeline_cls_name
+    if components.override_transformer_cls_name is not None:
+        kwargs["override_transformer_cls_name"] = components.override_transformer_cls_name
+    if components.text_encoder_weights is not None:
+        kwargs["override_text_encoder_safetensors"] = components.text_encoder_weights
+    if components.transformer_weights is not None:
+        kwargs["init_weights_from_safetensors"] = components.transformer_weights
+    if components.transformer_2_weights is not None:
+        kwargs["init_weights_from_safetensors_2"] = components.transformer_2_weights
+
+    kwargs.update(deepcopy(normalized.pipeline.profile_overrides))
+    kwargs.update(deepcopy(normalized.pipeline.experimental))
+    return FastVideoArgs.from_kwargs(**kwargs)
+
+
+def normalize_generation_request(request: GenerationRequest | Mapping[str, Any], ) -> GenerationRequest:
+    normalized = (request if isinstance(request, GenerationRequest) else parse_config(GenerationRequest, request))
+
+    if hasattr(normalized, EXPLICIT_REQUEST_ATTR):
+        refresh_generation_request_raw(normalized)
+    else:
+        bind_generation_request_raw(normalized, _serialize_generation_request(normalized))
+    return normalized
+
+
+def legacy_generate_call_to_request(
+    prompt: str | None,
+    sampling_param: SamplingParam | None,
+    *,
+    mouse_cond: Any | None = None,
+    keyboard_cond: Any | None = None,
+    grid_sizes: Any | None = None,
+    legacy_kwargs: Mapping[str, Any] | None = None,
+) -> GenerationRequest:
+    raw = _sampling_param_to_request_raw(sampling_param)
+    if prompt is not None:
+        raw["prompt"] = prompt
+
+    for key, value in (legacy_kwargs or {}).items():
+        _apply_request_field(raw, key, value)
+
+    if mouse_cond is not None:
+        raw.setdefault("inputs", {})["mouse_cond"] = mouse_cond
+    if keyboard_cond is not None:
+        raw.setdefault("inputs", {})["keyboard_cond"] = keyboard_cond
+    if grid_sizes is not None:
+        raw.setdefault("inputs", {})["grid_sizes"] = grid_sizes
+
+    normalized = parse_config(GenerationRequest, raw)
+    bind_generation_request_raw(normalized, raw)
+    return normalized
+
+
+def request_to_sampling_param(
+    request: GenerationRequest,
+    *,
+    model_path: str,
+) -> SamplingParam:
+    if request.plan is not None:
+        raise NotImplementedError("GenerationRequest.plan is not wired into VideoGenerator yet")
+    if request.state is not None:
+        raise NotImplementedError("GenerationRequest.state is not wired into VideoGenerator yet")
+
+    sampling_param = SamplingParam.from_pretrained(model_path)
+    updates = _explicit_request_updates(request)
+
+    for key, value in updates.items():
+        if hasattr(sampling_param, key):
+            setattr(sampling_param, key, deepcopy(value))
+        elif key in _REQUEST_PIPELINE_OVERRIDE_FIELDS or _is_supported_as_default_only(key, value):
+            continue
+        else:
+            raise ValueError(f"Request field {key!r} is not supported by sampling params for {model_path}")
+
+    sampling_param.__post_init__()
+    sampling_param.check_sampling_param()
+    return sampling_param
+
+
+def expand_request_prompt_batch(request: GenerationRequest, ) -> list[GenerationRequest]:
+    if not isinstance(request.prompt, list):
+        return [request]
+
+    requests: list[GenerationRequest] = []
+    for index, prompt in enumerate(request.prompt):
+        single_request = deepcopy(request)
+        single_request.prompt = prompt
+        _fan_out_batched_input_value(request, single_request, "image_path", index)
+        _fan_out_batched_input_value(request, single_request, "video_path", index)
+        _fan_out_explicit_request_metadata(request, single_request, index, prompt)
+        requests.append(single_request)
+    return requests
+
+
+def _looks_like_run_or_serve_config(raw: Mapping[str, Any]) -> bool:
+    return isinstance(raw.get("generator"), Mapping)
+
+
+def _normalize_overrides(overrides: list[str] | Mapping[str, Any] | None, ) -> dict[str, Any] | None:
+    if not overrides:
+        return None
+    if isinstance(overrides, list):
+        return parse_cli_overrides(overrides)
+    return dict(overrides)
+
+
+def _sampling_param_to_request_raw(sampling_param: SamplingParam | None, ) -> dict[str, Any]:
+    if sampling_param is None:
+        return {}
+
+    raw: dict[str, Any] = {}
+    for key, value in shallow_asdict(sampling_param).items():
+        if key == "prompt":
+            continue
+        _apply_request_field(raw, key, deepcopy(value))
+    return raw
+
+
+def _apply_request_field(
+    raw: dict[str, Any],
+    key: str,
+    value: Any,
+) -> None:
+    key = _LEGACY_REQUEST_ALIASES.get(key, key)
+    if key == "negative_prompt":
+        raw["negative_prompt"] = value
+        return
+    if key in _INPUT_FIELD_NAMES:
+        raw.setdefault("inputs", {})[key] = value
+        return
+    if key in _SAMPLING_FIELD_NAMES:
+        raw.setdefault("sampling", {})[key] = value
+        return
+    if key in _RUNTIME_FIELD_NAMES:
+        raw.setdefault("runtime", {})[key] = value
+        return
+    if key in _OUTPUT_FIELD_NAMES:
+        raw.setdefault("output", {})[key] = value
+        return
+    raw.setdefault("extensions", {})[key] = value
+
+
+def request_to_pipeline_overrides(request: GenerationRequest) -> dict[str, Any]:
+    overrides: dict[str, Any] = {}
+    for key, value in _explicit_request_updates(request).items():
+        if key in _REQUEST_PIPELINE_OVERRIDE_FIELDS:
+            overrides[key] = deepcopy(value)
+    return overrides
+
+
+def _explicit_request_updates(request: GenerationRequest) -> dict[str, Any]:
+    raw = getattr(request, EXPLICIT_REQUEST_ATTR, None)
+    if raw is None:
+        raw = _serialize_generation_request(request)
+
+    return _extract_request_updates(raw)
+
+
+def _extract_request_updates(raw: Mapping[str, Any]) -> dict[str, Any]:
+    updates: dict[str, Any] = {}
+    if "negative_prompt" in raw:
+        updates["negative_prompt"] = deepcopy(raw["negative_prompt"])
+
+    for section_name in ("inputs", "sampling", "runtime", "output"):
+        section = raw.get(section_name)
+        if not isinstance(section, Mapping):
+            continue
+        for key, value in section.items():
+            updates[key] = deepcopy(value)
+
+    stage_overrides = raw.get("stage_overrides")
+    if stage_overrides:
+        updates.update(_flatten_stage_overrides(stage_overrides))
+
+    extensions = raw.get("extensions")
+    if isinstance(extensions, Mapping):
+        for key, value in extensions.items():
+            updates[key] = deepcopy(value)
+
+    return updates
+
+
+def _flatten_stage_overrides(stage_overrides: Any) -> dict[str, Any]:
+    if not isinstance(stage_overrides, Mapping):
+        raise ValueError("GenerationRequest.stage_overrides must be a mapping")
+
+    flattened: dict[str, Any] = {}
+    for stage_name, overrides in stage_overrides.items():
+        if not isinstance(overrides, Mapping):
+            raise ValueError(f"GenerationRequest.stage_overrides.{stage_name} must be a mapping")
+        for key, value in overrides.items():
+            if key in flattened and flattened[key] != value:
+                raise ValueError(f"Conflicting stage override for {key!r} across stages")
+            flattened[key] = deepcopy(value)
+    return flattened
+
+
+def _serialize_generation_request(request: GenerationRequest) -> dict[str, Any]:
+    return deepcopy(config_to_dict(request))
+
+
+def _fan_out_batched_input_value(
+    source_request: GenerationRequest,
+    target_request: GenerationRequest,
+    field_name: str,
+    index: int,
+) -> None:
+    value = getattr(source_request.inputs, field_name)
+    if not isinstance(value, list):
+        return
+    _validate_batched_input_length(source_request.prompt, value, field_name)
+    setattr(target_request.inputs, field_name, deepcopy(value[index]))
+
+
+def _fan_out_explicit_request_metadata(
+    source_request: GenerationRequest,
+    target_request: GenerationRequest,
+    index: int,
+    prompt: str,
+) -> None:
+    raw = getattr(source_request, EXPLICIT_REQUEST_ATTR, None)
+    if raw is None:
+        return
+
+    raw = deepcopy(raw)
+    raw["prompt"] = prompt
+    inputs = raw.get("inputs")
+    if isinstance(inputs, dict):
+        for field_name in ("image_path", "video_path"):
+            value = inputs.get(field_name)
+            if isinstance(value, list):
+                _validate_batched_input_length(source_request.prompt, value, field_name)
+                inputs[field_name] = deepcopy(value[index])
+
+    setattr(target_request, EXPLICIT_REQUEST_ATTR, raw)
+
+
+def _validate_batched_input_length(
+    prompts: str | list[str] | None,
+    values: list[Any],
+    field_name: str,
+) -> None:
+    if not isinstance(prompts, list):
+        return
+    if len(values) != len(prompts):
+        raise ValueError(f"GenerationRequest.inputs.{field_name} must have the same length as request.prompt")
+
+
+def _is_supported_as_default_only(key: str, value: Any) -> bool:
+    default_value = _DEFAULT_REQUEST_UPDATES.get(key, _MISSING)
+    return default_value is not _MISSING and _values_equal(value, default_value)
+
+
+def _collect_non_default_fields(
+    value: Any,
+    default: Any,
+) -> dict[str, Any]:
+    if not (is_dataclass(value) and is_dataclass(default)):
+        return {}
+
+    result: dict[str, Any] = {}
+    for field in fields(value):
+        current = getattr(value, field.name)
+        default_value = getattr(default, field.name)
+        if is_dataclass(current) and is_dataclass(default_value):
+            nested = _collect_non_default_fields(current, default_value)
+            if nested:
+                result[field.name] = nested
+            continue
+        if not _values_equal(current, default_value):
+            result[field.name] = deepcopy(current)
+    return result
+
+
+def _values_equal(left: Any, right: Any) -> bool:
+    if left is right:
+        return True
+    try:
+        return bool(left == right)
+    except Exception:
+        return False
+
+
+_DEFAULT_REQUEST_UPDATES = _extract_request_updates(config_to_dict(GenerationRequest()))
+
+__all__ = [
+    "generator_config_to_fastvideo_args",
+    "legacy_from_pretrained_to_config",
+    "legacy_generate_call_to_request",
+    "load_generator_config_from_file",
+    "normalize_generation_request",
+    "normalize_generator_config",
+    "request_to_pipeline_overrides",
+    "request_to_sampling_param",
+]
@@ -0,0 +1,16 @@
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+
+class ConfigValidationError(ValueError):
+    """Validation error that keeps track of the nested config path."""
+
+    def __init__(self, path: str, message: str):
+        self.path = path
+        self.message = message
+        super().__init__(str(self))
+
+    def __str__(self) -> str:
+        if self.path:
+            return f"{self.path}: {self.message}"
+        return self.message
@@ -0,0 +1,101 @@
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+from copy import deepcopy
+from typing import Any
+from collections.abc import Mapping
+
+import yaml
+
+from fastvideo.api.errors import ConfigValidationError
+
+
+def parse_cli_overrides(overrides: list[str]) -> dict[str, Any]:
+    """Parse ``--dotted.key value`` style overrides into a flat mapping."""
+    parsed: dict[str, Any] = {}
+    index = 0
+    while index < len(overrides):
+        token = overrides[index]
+        if not token.startswith("--"):
+            raise ValueError(f"Expected --dotted.key, got {token!r}")
+
+        key = token[2:]
+        if not key:
+            raise ValueError("Override key cannot be empty")
+
+        if "=" in key:
+            key, raw_value = key.split("=", 1)
+        else:
+            index += 1
+            if index >= len(overrides):
+                raise ValueError(f"Missing value for override {token!r}")
+            raw_value = overrides[index]
+
+        parsed[_normalize_override_key(key)] = _cast_override_value(raw_value)
+        index += 1
+
+    return parsed
+
+
+def apply_overrides(config: Mapping[str, Any], overrides: Mapping[str, Any]) -> dict[str, Any]:
+    """Return a copy of ``config`` with dotted-key overrides applied."""
+    merged = deepcopy(dict(config))
+    for dotted_key, value in overrides.items():
+        _apply_single_override(merged, dotted_key, value)
+    return merged
+
+
+def _apply_single_override(config: dict[str, Any], dotted_key: str, value: Any) -> None:
+    parts = dotted_key.split(".")
+    if not all(parts):
+        raise ValueError(f"Invalid override path {dotted_key!r}")
+
+    cursor = config
+    for depth, part in enumerate(parts[:-1]):
+        existing = cursor.get(part)
+        if existing is None:
+            existing = {}
+            cursor[part] = existing
+        elif not isinstance(existing, dict):
+            raise ConfigValidationError(
+                ".".join(parts[:depth + 1]),
+                "cannot apply nested override through a non-mapping value",
+            )
+        cursor = existing
+
+    cursor[parts[-1]] = value
+
+
+def _cast_override_value(raw: str) -> Any:
+    lowered = raw.lower()
+    if lowered == "true":
+        return True
+    if lowered == "false":
+        return False
+    if lowered in {"none", "null"}:
+        return None
+
+    try:
+        return int(raw)
+    except ValueError:
+        pass
+
+    try:
+        return float(raw)
+    except ValueError:
+        pass
+
+    if raw.startswith("[") or raw.startswith("{"):
+        try:
+            return yaml.safe_load(raw)
+        except yaml.YAMLError:
+            pass
+
+    return raw
+
+
+def _normalize_override_key(key: str) -> str:
+    return key.replace("-", "_")
+
+
+__all__ = ["apply_overrides", "parse_cli_overrides"]
@@ -0,0 +1,332 @@
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+import dataclasses
+import json
+import types
+from pathlib import Path
+from collections.abc import Mapping
+from typing import Any, Literal, TypeVar, Union, get_args, get_origin, get_type_hints
+
+import yaml
+
+from fastvideo.api.errors import ConfigValidationError
+from fastvideo.api.overrides import apply_overrides, parse_cli_overrides
+from fastvideo.api.request_metadata import (
+    bind_generation_request_raw,
+    bind_run_config_raw,
+    bind_serve_config_raw,
+)
+from fastvideo.api.schema import GenerationRequest, RunConfig, ServeConfig
+
+T = TypeVar("T")
+_UNION_ORIGINS = {types.UnionType, Union}
+
+
+@dataclasses.dataclass(frozen=True)
+class _DataclassSpec:
+    cls: type[Any]
+    type_hints: dict[str, Any]
+    fields_by_name: dict[str, dataclasses.Field[Any]]
+
+
+def parse_config(config_type: type[T], raw: Mapping[str, Any] | T) -> T:
+    """Parse a nested mapping into a typed inference config object."""
+    if isinstance(raw, config_type):
+        return raw
+    if not isinstance(raw, Mapping):
+        raise ConfigValidationError("", f"expected mapping for {config_type.__name__}")
+    parsed = _SchemaParser().parse_dataclass(config_type, raw, "")
+    if config_type is GenerationRequest:
+        return bind_generation_request_raw(parsed, raw)
+    if config_type is RunConfig:
+        return bind_run_config_raw(parsed, raw)
+    if config_type is ServeConfig:
+        return bind_serve_config_raw(parsed, raw)
+    return parsed
+
+
+def config_to_dict(config: Any) -> Any:
+    """Serialize a typed config object into plain Python containers."""
+    if dataclasses.is_dataclass(config) and not isinstance(config, type):
+        return {field.name: config_to_dict(getattr(config, field.name)) for field in dataclasses.fields(config)}
+    if isinstance(config, list):
+        return [config_to_dict(item) for item in config]
+    if isinstance(config, dict):
+        return {key: config_to_dict(value) for key, value in config.items()}
+    return config
+
+
+def load_config(
+    config_type: type[T],
+    path: str | Path,
+    overrides: list[str] | Mapping[str, Any] | None = None,
+) -> T:
+    """Load a typed config object from YAML or JSON."""
+    raw = load_raw_config(path)
+    normalized_overrides = _normalize_overrides(overrides)
+    if normalized_overrides:
+        raw = apply_overrides(raw, normalized_overrides)
+    return parse_config(config_type, raw)
+
+
+def load_run_config(
+    path: str | Path,
+    overrides: list[str] | Mapping[str, Any] | None = None,
+) -> RunConfig:
+    return load_config(RunConfig, path, overrides)
+
+
+def load_serve_config(
+    path: str | Path,
+    overrides: list[str] | Mapping[str, Any] | None = None,
+) -> ServeConfig:
+    return load_config(ServeConfig, path, overrides)
+
+
+def load_raw_config(path: str | Path) -> dict[str, Any]:
+    config_path = Path(path)
+    if not config_path.exists():
+        raise FileNotFoundError(f"Config file not found: {config_path}")
+
+    with config_path.open(encoding="utf-8") as handle:
+        raw = _load_raw_mapping(handle, config_path)
+
+    if raw is None:
+        return {}
+    if not isinstance(raw, Mapping):
+        raise ConfigValidationError("", f"{config_path} must contain a top-level mapping")
+    return dict(raw)
+
+
+def _load_raw_mapping(handle: Any, config_path: Path) -> Any:
+    suffix = config_path.suffix.lower()
+    if suffix in {".yaml", ".yml"}:
+        return yaml.safe_load(handle)
+    if suffix == ".json":
+        return json.load(handle)
+    raise ValueError(f"Unsupported config file format: {config_path}")
+
+
+def _normalize_overrides(overrides: list[str] | Mapping[str, Any] | None, ) -> dict[str, Any] | None:
+    if not overrides:
+        return None
+    if isinstance(overrides, list):
+        return parse_cli_overrides(overrides)
+    return dict(overrides)
+
+
+class _SchemaParser:
+
+    def parse_dataclass(
+        self,
+        config_type: type[T],
+        raw: Mapping[str, Any],
+        path: str,
+    ) -> T:
+        if not isinstance(raw, Mapping):
+            raise ConfigValidationError(path, f"expected mapping for {config_type.__name__}")
+
+        spec = _get_dataclass_spec(config_type)
+        self._validate_keys(raw, spec, path)
+
+        values: dict[str, Any] = {}
+        for name, field in spec.fields_by_name.items():
+            field_path = _join_path(path, name)
+            if name in raw:
+                values[name] = self.parse_value(spec.type_hints[name], raw[name], field_path)
+                continue
+            if _field_is_required(field):
+                raise ConfigValidationError(field_path, "missing required field")
+
+        return config_type(**values)
+
+    def parse_value(self, annotation: Any, value: Any, path: str) -> Any:
+        if annotation is Any:
+            return value
+
+        origin = get_origin(annotation)
+        if origin in _UNION_ORIGINS:
+            return self._parse_union(annotation, value, path)
+        if origin is Literal:
+            return self._parse_literal(annotation, value, path)
+        if origin is list:
+            return self._parse_list(annotation, value, path)
+        if origin is dict:
+            return self._parse_dict(annotation, value, path)
+        if origin is tuple:
+            return self._parse_tuple(annotation, value, path)
+        if isinstance(annotation, type) and dataclasses.is_dataclass(annotation):
+            return self.parse_dataclass(annotation, value, path)
+
+        scalar_parser = _SCALAR_PARSERS.get(annotation)
+        if scalar_parser is not None:
+            return scalar_parser(value, path)
+
+        return self._parse_instance(annotation, value, path)
+
+    def _validate_keys(
+        self,
+        raw: Mapping[str, Any],
+        spec: _DataclassSpec,
+        path: str,
+    ) -> None:
+        for key in raw:
+            if not isinstance(key, str):
+                raise ConfigValidationError(path, "expected mapping keys to be strings")
+            if key not in spec.fields_by_name:
+                raise ConfigValidationError(_join_path(path, key), "unknown field")
+
+    def _parse_union(self, annotation: Any, value: Any, path: str) -> Any:
+        candidates = [candidate for candidate in get_args(annotation) if candidate is not type(None)]
+        if value is None and len(candidates) != len(get_args(annotation)):
+            return None
+        if len(candidates) == 1:
+            return self.parse_value(candidates[0], value, path)
+
+        errors: list[str] = []
+        for candidate in candidates:
+            try:
+                return self.parse_value(candidate, value, path)
+            except ConfigValidationError as exc:
+                errors.append(exc.message)
+
+        expected = ", ".join(_type_name(candidate) for candidate in candidates)
+        detail = errors[0] if errors else f"expected one of ({expected})"
+        raise ConfigValidationError(path, detail)
+
+    def _parse_literal(self, annotation: Any, value: Any, path: str) -> Any:
+        allowed = get_args(annotation)
+        if value not in allowed:
+            raise ConfigValidationError(path, f"expected one of {sorted(allowed)!r}")
+        return value
+
+    def _parse_list(self, annotation: Any, value: Any, path: str) -> list[Any]:
+        if not isinstance(value, list):
+            raise ConfigValidationError(path, "expected list")
+        item_type = get_args(annotation)[0] if get_args(annotation) else Any
+        return [self.parse_value(item_type, item, f"{path}[{index}]") for index, item in enumerate(value)]
+
+    def _parse_dict(self, annotation: Any, value: Any, path: str) -> dict[Any, Any]:
+        if not isinstance(value, Mapping):
+            raise ConfigValidationError(path, "expected mapping")
+
+        key_type, value_type = (get_args(annotation) + (Any, Any))[:2]
+        parsed: dict[Any, Any] = {}
+        for key, item in value.items():
+            parsed_key = self._parse_dict_key(key_type, key, path)
+            item_path = _join_path(path, str(key))
+            parsed[parsed_key] = self.parse_value(value_type, item, item_path)
+        return parsed
+
+    def _parse_tuple(self, annotation: Any, value: Any, path: str) -> tuple[Any, ...]:
+        if not isinstance(value, list | tuple):
+            raise ConfigValidationError(path, "expected tuple")
+
+        item_types = get_args(annotation)
+        if len(item_types) == 2 and item_types[1] is Ellipsis:
+            return tuple(self.parse_value(item_types[0], item, f"{path}[{index}]") for index, item in enumerate(value))
+
+        if len(value) != len(item_types):
+            raise ConfigValidationError(path, f"expected tuple of length {len(item_types)}")
+
+        return tuple(
+            self.parse_value(item_type, item, f"{path}[{index}]")
+            for index, (item_type, item) in enumerate(zip(item_types, value, strict=True)))
+
+    def _parse_dict_key(self, annotation: Any, value: Any, path: str) -> Any:
+        if annotation is Any:
+            return value
+        if annotation is str:
+            if not isinstance(value, str):
+                raise ConfigValidationError(path, "expected string dictionary keys")
+            return value
+        if annotation is int:
+            if not isinstance(value, int) or isinstance(value, bool):
+                raise ConfigValidationError(path, "expected integer dictionary keys")
+            return value
+        return value
+
+    def _parse_instance(self, annotation: Any, value: Any, path: str) -> Any:
+        if isinstance(annotation, type) and not isinstance(value, annotation):
+            raise ConfigValidationError(path, f"expected {annotation.__name__}")
+        return value
+
+
+def _parse_bool(value: Any, path: str) -> bool:
+    if type(value) is not bool:
+        raise ConfigValidationError(path, "expected bool")
+    return value
+
+
+def _parse_int(value: Any, path: str) -> int:
+    if not isinstance(value, int) or isinstance(value, bool):
+        raise ConfigValidationError(path, "expected int")
+    return value
+
+
+def _parse_float(value: Any, path: str) -> float:
+    if not isinstance(value, int | float) or isinstance(value, bool):
+        raise ConfigValidationError(path, "expected float")
+    return float(value)
+
+
+def _parse_str(value: Any, path: str) -> str:
+    if not isinstance(value, str):
+        raise ConfigValidationError(path, "expected str")
+    return value
+
+
+_SCALAR_PARSERS: dict[Any, Any] = {
+    bool: _parse_bool,
+    int: _parse_int,
+    float: _parse_float,
+    str: _parse_str,
+}
+
+
+def _field_is_required(field: dataclasses.Field[Any]) -> bool:
+    return (field.default is dataclasses.MISSING and field.default_factory is dataclasses.MISSING)
+
+
+def _get_dataclass_spec(config_type: type[Any]) -> _DataclassSpec:
+    spec = _DATACLASS_SPEC_CACHE.get(config_type)
+    if spec is not None:
+        return spec
+
+    spec = _DataclassSpec(
+        cls=config_type,
+        type_hints=get_type_hints(config_type),
+        fields_by_name={field.name: field
+                        for field in dataclasses.fields(config_type)},
+    )
+    _DATACLASS_SPEC_CACHE[config_type] = spec
+    return spec
+
+
+_DATACLASS_SPEC_CACHE: dict[type[Any], _DataclassSpec] = {}
+
+
+def _join_path(prefix: str, suffix: str) -> str:
+    if not prefix:
+        return suffix
+    return f"{prefix}.{suffix}"
+
+
+def _type_name(annotation: Any) -> str:
+    origin = get_origin(annotation)
+    if origin is not None:
+        return str(annotation)
+    if hasattr(annotation, "__name__"):
+        return annotation.__name__
+    return str(annotation)
+
+
+__all__ = [
+    "config_to_dict",
+    "load_config",
+    "load_raw_config",
+    "load_run_config",
+    "load_serve_config",
+    "parse_config",
+]
@@ -0,0 +1,247 @@
+# SPDX-License-Identifier: Apache-2.0
+"""Track which GenerationRequest fields the user explicitly provided.
+
+This module solves a specific problem: when translating a GenerationRequest into
+a legacy SamplingParam, we need to distinguish user-provided values (which
+should override model defaults) from schema defaults (which should NOT override
+model defaults).
+
+The approach:
+1. At bind time, store the original raw dict and a baseline snapshot.
+2. Patch __setattr__ on tracked dataclass types to record dirty field paths.
+3. At access time, do a lazy 3-way merge: raw + baseline + current state,
+   with dirty paths forcing inclusion even when current == baseline.
+"""
+from __future__ import annotations
+
+from collections.abc import Mapping
+from copy import deepcopy
+import dataclasses
+from typing import Any, cast
+from collections.abc import Callable
+
+from fastvideo.api.schema import (
+    ContinuationState,
+    GenerationPlan,
+    GenerationRequest,
+    InputConfig,
+    OutputConfig,
+    PlannedStage,
+    RequestRuntimeConfig,
+    RunConfig,
+    SamplingConfig,
+    ServeConfig,
+)
+
+EXPLICIT_REQUEST_ATTR = "_fastvideo_explicit_request"
+ORIGINAL_REQUEST_STATE_ATTR = "_fastvideo_original_request_state"
+_TRACKING_ROOT_ATTR = "_fastvideo_request_tracking_root"
+_TRACKING_PATH_ATTR = "_fastvideo_request_tracking_path"
+_TRACKING_PATCHED_ATTR = "_fastvideo_request_tracking_patched"
+_DIRTY_PATHS_ATTR = "_fastvideo_dirty_paths"
+_TRACKED_REQUEST_TYPES = (
+    GenerationRequest,
+    InputConfig,
+    SamplingConfig,
+    RequestRuntimeConfig,
+    OutputConfig,
+    ContinuationState,
+    PlannedStage,
+    GenerationPlan,
+)
+
+
+def bind_generation_request_raw(
+    request: GenerationRequest,
+    raw: Mapping[str, Any] | None,
+) -> GenerationRequest:
+    _ensure_request_tracking()
+    # Disable dirty tracking during bind so tree walk doesn't record paths.
+    object.__setattr__(request, _DIRTY_PATHS_ATTR, None)
+    object.__setattr__(request, EXPLICIT_REQUEST_ATTR, deepcopy(dict(raw or {})))
+    object.__setattr__(request, ORIGINAL_REQUEST_STATE_ATTR, _serialize_config(request))
+    _set_tracking_roots(request, request, "")
+    # Enable dirty tracking.
+    object.__setattr__(request, _DIRTY_PATHS_ATTR, set())
+    return request
+
+
+def bind_run_config_raw(
+    config: RunConfig,
+    raw: Mapping[str, Any],
+) -> RunConfig:
+    request_raw = raw.get("request")
+    if isinstance(request_raw, Mapping):
+        bind_generation_request_raw(config.request, request_raw)
+    return config
+
+
+def bind_serve_config_raw(
+    config: ServeConfig,
+    raw: Mapping[str, Any],
+) -> ServeConfig:
+    default_request_raw = raw.get("default_request")
+    if isinstance(default_request_raw, Mapping):
+        bind_generation_request_raw(config.default_request, default_request_raw)
+    elif "default_request" not in raw:
+        bind_generation_request_raw(config.default_request, {})
+    return config
+
+
+def refresh_generation_request_raw(request: GenerationRequest, ) -> dict[str, Any] | None:
+    raw = getattr(request, EXPLICIT_REQUEST_ATTR, None)
+    baseline = getattr(request, ORIGINAL_REQUEST_STATE_ATTR, None)
+    if not isinstance(raw, Mapping) or not isinstance(baseline, Mapping):
+        return None
+
+    dirty = getattr(request, _DIRTY_PATHS_ATTR, None) or frozenset()
+    current = _serialize_config(request)
+    merged = deepcopy(dict(raw))
+    _merge_request_mutations(merged, dict(baseline), current, dirty)
+
+    object.__setattr__(request, EXPLICIT_REQUEST_ATTR, merged)
+    object.__setattr__(request, ORIGINAL_REQUEST_STATE_ATTR, current)
+    object.__setattr__(request, _DIRTY_PATHS_ATTR, set())
+    return merged
+
+
+# ---------------------------------------------------------------------------
+# 3-way merge: raw + baseline + current, with dirty-path forcing
+# ---------------------------------------------------------------------------
+
+_MISSING = object()
+
+
+def _merge_request_mutations(
+    merged: dict[str, Any],
+    baseline: Mapping[str, Any],
+    current: Mapping[str, Any],
+    dirty: frozenset[str] | set[str],
+    path_prefix: str = "",
+    force_dirty: bool = False,
+) -> None:
+    # Remove keys that were deleted from the current state.
+    for key in set(merged) | set(baseline):
+        if key not in current:
+            merged.pop(key, None)
+
+    for key in current:
+        current_path = f"{path_prefix}.{key}" if path_prefix else key
+        current_value = current[key]
+        baseline_value = baseline.get(key, _MISSING)
+        merged_value = merged.get(key, _MISSING)
+
+        # If this exact path was dirtied (e.g. whole section replaced),
+        # propagate to all children.
+        child_force = force_dirty or current_path in dirty
+
+        # Recurse into nested mappings.
+        if isinstance(current_value, Mapping) and isinstance(baseline_value, Mapping):
+            nested = (deepcopy(dict(merged_value)) if isinstance(merged_value, Mapping) else {})
+            _merge_request_mutations(
+                nested,
+                baseline_value,
+                current_value,
+                dirty,
+                current_path,
+                child_force,
+            )
+            if nested:
+                merged[key] = nested
+            else:
+                merged.pop(key, None)
+            continue
+
+        # A field is explicitly set if:
+        # - it's new (not in baseline),
+        # - it changed from baseline,
+        # - its path was touched by __setattr__ (dirty), or
+        # - an ancestor path was dirty (whole section replaced).
+        is_dirty = child_force or current_path in dirty
+        if baseline_value is _MISSING or current_value != baseline_value or is_dirty:
+            merged[key] = deepcopy(current_value)
+
+
+# ---------------------------------------------------------------------------
+# __setattr__ patching for dirty-path recording
+# ---------------------------------------------------------------------------
+
+
+def _ensure_request_tracking() -> None:
+    for config_type in _TRACKED_REQUEST_TYPES:
+        _patch_tracking_setattr(config_type)
+
+
+def _patch_tracking_setattr(config_type: type[Any]) -> None:
+    if getattr(config_type, _TRACKING_PATCHED_ATTR, False):
+        return
+
+    original_setattr = cast(
+        Callable[[Any, str, Any], None],
+        config_type.__setattr__,
+    )
+    field_names = {field.name for field in dataclasses.fields(config_type)}
+
+    def _tracking_setattr(self: Any, name: str, value: Any) -> None:
+        if name.startswith("_fastvideo_") or name not in field_names:
+            original_setattr(self, name, value)
+            return
+
+        root = getattr(self, _TRACKING_ROOT_ATTR, None)
+        if root is not None:
+            dirty = getattr(root, _DIRTY_PATHS_ATTR, None)
+            if isinstance(dirty, set):
+                prefix = getattr(self, _TRACKING_PATH_ATTR, "")
+                path = f"{prefix}.{name}" if prefix else name
+                dirty.add(path)
+
+        original_setattr(self, name, value)
+
+    type.__setattr__(config_type, "__setattr__", _tracking_setattr)
+    setattr(config_type, _TRACKING_PATCHED_ATTR, True)
+
+
+# ---------------------------------------------------------------------------
+# Tree walk to set tracking root/path on nested dataclasses
+# ---------------------------------------------------------------------------
+
+
+def _set_tracking_roots(
+    root: GenerationRequest,
+    obj: Any,
+    prefix: str,
+) -> None:
+    if not dataclasses.is_dataclass(obj) or isinstance(obj, type):
+        return
+    object.__setattr__(obj, _TRACKING_ROOT_ATTR, root)
+    object.__setattr__(obj, _TRACKING_PATH_ATTR, prefix)
+    for field in dataclasses.fields(obj):
+        child = getattr(obj, field.name)
+        child_path = f"{prefix}.{field.name}" if prefix else field.name
+        if dataclasses.is_dataclass(child) and not isinstance(child, type):
+            _set_tracking_roots(root, child, child_path)
+
+
+# ---------------------------------------------------------------------------
+# Serialization helper
+# ---------------------------------------------------------------------------
+
+
+def _serialize_config(config: Any) -> Any:
+    if dataclasses.is_dataclass(config) and not isinstance(config, type):
+        return {field.name: _serialize_config(getattr(config, field.name)) for field in dataclasses.fields(config)}
+    if isinstance(config, list):
+        return [_serialize_config(item) for item in config]
+    if isinstance(config, dict):
+        return {key: _serialize_config(value) for key, value in config.items()}
+    return deepcopy(config)
+
+
+__all__ = [
+    "EXPLICIT_REQUEST_ATTR",
+    "ORIGINAL_REQUEST_STATE_ATTR",
+    "bind_generation_request_raw",
+    "bind_run_config_raw",
+    "bind_serve_config_raw",
+    "refresh_generation_request_raw",
+]
@@ -0,0 +1,101 @@
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Any
+from collections.abc import Mapping
+
+from fastvideo.api.schema import ContinuationState
+
+
+@dataclass
+class GenerationResult:
+    prompt: str | None = None
+    prompt_index: int | None = None
+    samples: Any | None = None
+    frames: Any | None = None
+    audio: Any | None = None
+    size: tuple[int, int, int] | None = None
+    generation_time: float | None = None
+    logging_info: Any | None = None
+    trajectory: Any | None = None
+    trajectory_timesteps: Any | None = None
+    trajectory_decoded: Any | None = None
+    video_path: str | None = None
+    peak_memory_mb: float | None = None
+    state: ContinuationState | None = None
+    extra: dict[str, Any] = field(default_factory=dict)
+
+    @classmethod
+    def from_legacy_result(
+        cls,
+        result: Mapping[str, Any],
+    ) -> GenerationResult:
+        prompt = result.get("prompt")
+        if prompt is None:
+            prompt = result.get("prompts")
+
+        extra = {
+            key: value
+            for key, value in result.items() if key not in {
+                "prompt",
+                "prompt_index",
+                "prompts",
+                "samples",
+                "frames",
+                "audio",
+                "size",
+                "generation_time",
+                "logging_info",
+                "trajectory",
+                "trajectory_timesteps",
+                "trajectory_decoded",
+                "video_path",
+                "peak_memory_mb",
+                "state",
+            }
+        }
+
+        return cls(
+            prompt=prompt,
+            prompt_index=result.get("prompt_index"),
+            samples=result.get("samples"),
+            frames=result.get("frames"),
+            audio=result.get("audio"),
+            size=result.get("size"),
+            generation_time=result.get("generation_time"),
+            logging_info=result.get("logging_info"),
+            trajectory=result.get("trajectory"),
+            trajectory_timesteps=result.get("trajectory_timesteps"),
+            trajectory_decoded=result.get("trajectory_decoded"),
+            video_path=result.get("video_path"),
+            peak_memory_mb=result.get("peak_memory_mb"),
+            state=result.get("state"),
+            extra=extra,
+        )
+
+    def to_legacy_dict(self) -> dict[str, Any]:
+        result = {
+            "prompts": self.prompt,
+            "samples": self.samples,
+            "frames": self.frames,
+            "audio": self.audio,
+            "size": self.size,
+            "generation_time": self.generation_time,
+            "logging_info": self.logging_info,
+            "trajectory": self.trajectory,
+            "trajectory_timesteps": self.trajectory_timesteps,
+            "trajectory_decoded": self.trajectory_decoded,
+            "video_path": self.video_path,
+            "peak_memory_mb": self.peak_memory_mb,
+        }
+        if self.prompt_index is not None:
+            result["prompt_index"] = self.prompt_index
+            result["prompt"] = self.prompt
+        if self.state is not None:
+            result["state"] = self.state
+        result.update(self.extra)
+        return result
+
+
+__all__ = ["GenerationResult"]
@@ -0,0 +1,210 @@
+# SPDX-License-Identifier: Apache-2.0
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from typing import Any, Literal
+
+
+@dataclass
+class ServerConfig:
+    host: str = "0.0.0.0"
+    port: int = 8000
+    output_dir: str = "outputs/"
+
+
+@dataclass
+class ParallelismConfig:
+    tp_size: int = -1
+    sp_size: int = -1
+    hsdp_replicate_dim: int = 1
+    hsdp_shard_dim: int = -1
+    dist_timeout: int | None = None
+
+
+@dataclass
+class OffloadConfig:
+    dit: bool = True
+    dit_layerwise: bool = True
+    text_encoder: bool = True
+    image_encoder: bool = True
+    vae: bool = True
+    pin_cpu_memory: bool = True
+
+
+@dataclass
+class CompileConfig:
+    enabled: bool = False
+    kwargs: dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class QuantizationConfig:
+    text_encoder_quant: str | None = None
+    transformer_quant: str | None = None
+
+
+@dataclass
+class EngineConfig:
+    num_gpus: int = 1
+    execution_backend: Literal["mp", "ray"] = "mp"
+    parallelism: ParallelismConfig = field(default_factory=ParallelismConfig)
+    offload: OffloadConfig = field(default_factory=OffloadConfig)
+    compile: CompileConfig = field(default_factory=CompileConfig)
+    enable_stage_verification: bool = True
+    use_fsdp_inference: bool = False
+    disable_autocast: bool = False
+    quantization: QuantizationConfig | None = None
+
+
+@dataclass
+class ComponentConfig:
+    config_root: str | None = None
+    pipeline_config_path: str | None = None
+    text_encoder_weights: str | None = None
+    transformer_weights: str | None = None
+    transformer_2_weights: str | None = None
+    vae_weights: str | None = None
+    upsampler_weights: str | None = None
+    lora_path: str | None = None
+    override_pipeline_cls_name: str | None = None
+    override_transformer_cls_name: str | None = None
+
+
+@dataclass
+class PipelineSelection:
+    workload_type: Literal["t2v", "i2v", "t2i", "i2i"] | None = None
+    profile: str | None = None
+    profile_version: str | None = None
+    components: ComponentConfig = field(default_factory=ComponentConfig)
+    profile_overrides: dict[str, Any] = field(default_factory=dict)
+    experimental: dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class GeneratorConfig:
+    model_path: str
+    revision: str | None = None
+    trust_remote_code: bool = False
+    engine: EngineConfig = field(default_factory=EngineConfig)
+    pipeline: PipelineSelection = field(default_factory=PipelineSelection)
+
+
+@dataclass
+class InputConfig:
+    prompt_path: str | None = None
+    image_path: str | list[str] | None = None
+    video_path: str | list[str] | None = None
+    pil_image: Any | None = None
+    pose: str | None = None
+    mouse_cond: Any | None = None
+    keyboard_cond: Any | None = None
+    grid_sizes: Any | None = None
+    c2ws_plucker_emb: Any | None = None
+    refine_from: str | None = None
+    stage1_video: Any | None = None
+
+
+@dataclass
+class SamplingConfig:
+    num_videos_per_prompt: int = 1
+    seed: int = 1024
+    num_frames: int = 125
+    height: int = 720
+    width: int = 1280
+    height_sr: int = 1072
+    width_sr: int = 1920
+    fps: int = 24
+    num_inference_steps: int = 50
+    num_inference_steps_sr: int = 50
+    guidance_scale: float = 1.0
+    guidance_scale_2: float | None = None
+    guidance_rescale: float = 0.0
+    true_cfg_scale: float | None = None
+    boundary_ratio: float | None = None
+    sigmas: list[float] | None = None
+
+
+@dataclass
+class RequestRuntimeConfig:
+    enable_teacache: bool = False
+    return_trajectory_latents: bool = False
+    return_trajectory_decoded: bool = False
+
+
+@dataclass
+class OutputConfig:
+    output_path: str = "outputs/"
+    output_video_name: str | None = None
+    save_video: bool = True
+    return_frames: bool = True
+    return_state: bool = False
+
+
+@dataclass
+class ContinuationState:
+    kind: str
+    payload: dict[str, Any]
+
+
+@dataclass
+class PlannedStage:
+    name: str
+    kind: str
+    source: str | None = None
+    overrides: dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class GenerationPlan:
+    stages: list[PlannedStage]
+    final_stage: str | None = None
+
+
+@dataclass
+class GenerationRequest:
+    prompt: str | list[str] | None = None
+    negative_prompt: str | None = None
+    inputs: InputConfig = field(default_factory=InputConfig)
+    sampling: SamplingConfig = field(default_factory=SamplingConfig)
+    runtime: RequestRuntimeConfig = field(default_factory=RequestRuntimeConfig)
+    output: OutputConfig = field(default_factory=OutputConfig)
+    stage_overrides: dict[str, Any] = field(default_factory=dict)
+    state: ContinuationState | None = None
+    plan: GenerationPlan | None = None
+    extensions: dict[str, Any] = field(default_factory=dict)
+
+
+@dataclass
+class RunConfig:
+    generator: GeneratorConfig
+    request: GenerationRequest
+
+
+@dataclass
+class ServeConfig:
+    generator: GeneratorConfig
+    server: ServerConfig = field(default_factory=ServerConfig)
+    default_request: GenerationRequest = field(default_factory=GenerationRequest)
+
+
+__all__ = [
+    "CompileConfig",
+    "ComponentConfig",
+    "ContinuationState",
+    "EngineConfig",
+    "GenerationPlan",
+    "GenerationRequest",
+    "GeneratorConfig",
+    "InputConfig",
+    "OffloadConfig",
+    "OutputConfig",
+    "ParallelismConfig",
+    "PipelineSelection",
+    "PlannedStage",
+    "QuantizationConfig",
+    "RequestRuntimeConfig",
+    "RunConfig",
+    "SamplingConfig",
+    "ServeConfig",
+    "ServerConfig",
+]
@@ -1,10 +1,7 @@
 # SPDX-License-Identifier: Apache-2.0

-from fastvideo.attention.backends.abstract import (AttentionBackend,
-                                                   AttentionMetadata,
-                                                   AttentionMetadataBuilder)
-from fastvideo.attention.layer import (DistributedAttention,
-                                       DistributedAttention_VSA, LocalAttention)
+from fastvideo.attention.backends.abstract import (AttentionBackend, AttentionMetadata, AttentionMetadataBuilder)
+from fastvideo.attention.layer import (DistributedAttention, DistributedAttention_VSA, LocalAttention)
 from fastvideo.attention.selector import get_attn_backend

 __all__ = [
@@ -54,17 +54,13 @@ class AttentionMetadata:
    # Current step of diffusion process
    current_timestep: int

-    def asdict_zerocopy(self,
-                        skip_fields: set[str] | None = None) -> dict[str, Any]:
+    def asdict_zerocopy(self, skip_fields: set[str] | None = None) -> dict[str, Any]:
        """Similar to dataclasses.asdict, but avoids deepcopying."""
        if skip_fields is None:
            skip_fields = set()
        # Note that if we add dataclasses as fields, they will need
        # similar handling.
-        return {
-            field.name: getattr(self, field.name)
-            for field in fields(self) if field.name not in skip_fields
-        }
+        return {field.name: getattr(self, field.name) for field in fields(self) if field.name not in skip_fields}


 T = TypeVar("T", bound=AttentionMetadata)
@@ -125,8 +121,7 @@ class AttentionImpl(ABC, Generic[T]):
    ) -> None:
        raise NotImplementedError

-    def preprocess_qkv(self, qkv: torch.Tensor,
-                       attn_metadata: T) -> torch.Tensor:
+    def preprocess_qkv(self, qkv: torch.Tensor, attn_metadata: T) -> torch.Tensor:
        """Preprocess QKV tensor before performing attention operation.

        Default implementation returns the tensor unchanged.
@@ -0,0 +1,738 @@
+# SPDX-License-Identifier: Apache-2.0
+"""
+Bidirectional Sparse Attention (BSA) backend for FastVideo.
+
+Pure-PyTorch reference implementation from:
+"Bidirectional Sparse Attention for Faster Video Diffusion Training"
+(arXiv:2509.01085)
+
+BSA sparsifies both queries (pruning redundant tokens per block) and
+key-value pairs (keeping only relevant KV blocks per query block).
+
+This is a training-free inference backend: it works with any model
+trained with full attention by applying BSA sparsity at inference time.
+"""
+
+import functools
+import math
+from dataclasses import dataclass
+from typing import Any
+
+import torch
+import torch.nn.functional as F
+
+from fastvideo.attention.backends.abstract import (
+    AttentionBackend,
+    AttentionImpl,
+    AttentionMetadata,
+    AttentionMetadataBuilder,
+)
+from fastvideo.distributed import get_sp_group
+from fastvideo.logger import init_logger
+
+try:
+    from fastvideo.attention.utils.flash_attn_no_pad import (
+        flash_attn_varlen_func_impl, )
+    FLASH_ATTN_AVAILABLE = True
+except ImportError:
+    try:
+        from flash_attn import flash_attn_varlen_func as flash_attn_varlen_func_impl
+        FLASH_ATTN_AVAILABLE = True
+    except ImportError:
+        FLASH_ATTN_AVAILABLE = False
+
+logger = init_logger(__name__)
+
+BSA_TILE_SIZE = (4, 4, 4)
+
+# ---------------------------------------------------------------------------
+# Cached index helpers (same pattern as VSA)
+# ---------------------------------------------------------------------------
+
+
+@functools.lru_cache(maxsize=10)
+def get_tile_partition_indices(
+    dit_seq_shape: tuple[int, int, int],
+    tile_size: tuple[int, int, int],
+    device: torch.device,
+) -> torch.LongTensor:
+    """Map raster-order tokens to tile-contiguous order."""
+    T, H, W = dit_seq_shape
+    ts, hs, ws = tile_size
+    indices = torch.arange(T * H * W, device=device, dtype=torch.long).reshape(T, H, W)
+    ls = []
+    for t in range(math.ceil(T / ts)):
+        for h in range(math.ceil(H / hs)):
+            for w in range(math.ceil(W / ws)):
+                ls.append(indices[
+                    t * ts:min(t * ts + ts, T),
+                    h * hs:min(h * hs + hs, H),
+                    w * ws:min(w * ws + ws, W),
+                ].flatten())
+    return torch.cat(ls, dim=0)
+
+
+@functools.lru_cache(maxsize=10)
+def get_reverse_tile_partition_indices(
+    dit_seq_shape: tuple[int, int, int],
+    tile_size: tuple[int, int, int],
+    device: torch.device,
+) -> torch.LongTensor:
+    """Inverse mapping: tile-contiguous order back to raster order."""
+    return torch.argsort(get_tile_partition_indices(dit_seq_shape, tile_size, device))
+
+
+# ---------------------------------------------------------------------------
+# BSA core operations
+# ---------------------------------------------------------------------------
+
+
+def _prune_queries(
+    q_blocks: torch.Tensor,
+    keep_ratio: float,
+) -> tuple[torch.Tensor, torch.Tensor, int]:
+    """
+    Prune redundant query tokens within each block.
+
+    Scores tokens by cosine similarity to the block center.
+    Keeps the LEAST similar (most informative) tokens.
+
+    Args:
+        q_blocks: [B, N_heads, N_blocks, block_size, D]
+        keep_ratio: fraction of tokens to keep
+
+    Returns:
+        sparse_q: [B, N_heads, N_blocks, keep_size, D]
+        keep_indices: [B, N_heads, N_blocks, keep_size]
+        keep_size: int
+    """
+    B, H, N, S, D = q_blocks.shape
+    keep_size = max(1, int(S * keep_ratio))
+
+    if keep_size >= S:
+        idx = torch.arange(S, device=q_blocks.device)
+        idx = idx.view(1, 1, 1, S).expand(B, H, N, S)
+        return q_blocks, idx, S
+
+    center_idx = S // 2
+    center = q_blocks[:, :, :, center_idx:center_idx + 1, :]
+
+    q_norm = F.normalize(q_blocks, dim=-1)
+    c_norm = F.normalize(center, dim=-1)
+    similarity = (q_norm * c_norm).sum(dim=-1)  # [B, H, N, S]
+
+    # lowest similarity = most distinctive = keep
+    _, indices = similarity.topk(keep_size, dim=-1, largest=False)
+    indices, _ = indices.sort(dim=-1)
+
+    idx_expand = indices.unsqueeze(-1).expand(-1, -1, -1, -1, D)
+    sparse_q = torch.gather(q_blocks, 3, idx_expand)
+
+    return sparse_q, indices, keep_size
+
+
+def _select_kv_blocks(
+    sparse_q: torch.Tensor,
+    k_blocks: torch.Tensor,
+    cumulative_threshold: float,
+    min_kv_blocks: int,
+) -> torch.Tensor:
+    """
+    Dynamically select KV blocks for each query block.
+
+    Mean-pools to block level, computes block attention scores,
+    admits blocks in descending order until cumulative mass
+    exceeds threshold.
+
+    Args:
+        sparse_q: [B, H, N, Sq, D]
+        k_blocks:  [B, H, N, Sk, D]
+        cumulative_threshold: e.g. 0.9
+        min_kv_blocks: minimum blocks to keep
+
+    Returns:
+        kv_mask: [B, H, N, N] boolean
+    """
+    B, H, N, _, D = sparse_q.shape
+
+    q_repr = sparse_q.mean(dim=3)
+    k_repr = k_blocks.mean(dim=3)
+
+    scores = torch.matmul(q_repr, k_repr.transpose(-1, -2)) / (D**0.5)
+    block_attn = F.softmax(scores, dim=-1)
+
+    sorted_attn, sorted_idx = block_attn.sort(dim=-1, descending=True)
+    cumsum = sorted_attn.cumsum(dim=-1)
+
+    keep_sorted = torch.ones_like(cumsum, dtype=torch.bool)
+    keep_sorted[..., 1:] = cumsum[..., :-1] < cumulative_threshold
+
+    min_mask = torch.zeros_like(keep_sorted)
+    min_mask[..., :min(min_kv_blocks, N)] = True
+    keep_sorted = keep_sorted | min_mask
+
+    kv_mask = torch.zeros_like(block_attn, dtype=torch.bool)
+    kv_mask.scatter_(-1, sorted_idx, keep_sorted)
+
+    return kv_mask
+
+
+def _compute_sparse_attention(
+    sparse_q: torch.Tensor,
+    k_blocks: torch.Tensor,
+    v_blocks: torch.Tensor,
+    kv_mask: torch.Tensor,
+) -> torch.Tensor:
+    """
+    Compute attention for each query block against selected KV blocks.
+
+    Handles per-batch and per-head KV masks correctly.
+    Uses flash_attn_varlen_func when available on GPU.
+    Falls back to pure-PyTorch reference on CPU.
+
+    Args:
+        sparse_q: [B, H, N, Sq, D]
+        k_blocks:  [B, H, N, Sk, D]
+        v_blocks:  [B, H, N, Sk, D]
+        kv_mask:   [B, H, N, N] boolean (per-batch, per-head)
+
+    Returns:
+        output: [B, H, N, Sq, D]
+    """
+    if FLASH_ATTN_AVAILABLE and sparse_q.is_cuda:
+        return _compute_sparse_attention_flash(sparse_q, k_blocks, v_blocks, kv_mask)
+    else:
+        return _compute_sparse_attention_reference(sparse_q, k_blocks, v_blocks, kv_mask)
+
+
+def _compute_sparse_attention_reference(
+    sparse_q: torch.Tensor,
+    k_blocks: torch.Tensor,
+    v_blocks: torch.Tensor,
+    kv_mask: torch.Tensor,
+) -> torch.Tensor:
+    """Pure-PyTorch fallback with per-batch, per-head mask support."""
+    B, H, N, Sq, D = sparse_q.shape
+    output = torch.zeros_like(sparse_q)
+
+    for b in range(B):
+        for h in range(H):
+            for qb in range(N):
+                selected = kv_mask[b, h, qb]  # [N] boolean
+                sel_idx = selected.nonzero(as_tuple=True)[0]
+
+                if sel_idx.shape[0] == 0:
+                    continue
+
+                # [num_sel * Sk, D]
+                sel_k = k_blocks[b, h, sel_idx].reshape(-1, D)
+                sel_v = v_blocks[b, h, sel_idx].reshape(-1, D)
+
+                q = sparse_q[b, h, qb]  # [Sq, D]
+                scores = torch.matmul(q, sel_k.transpose(-1, -2)) / (D**0.5)
+                weights = F.softmax(scores, dim=-1)
+                output[b, h, qb] = torch.matmul(weights, sel_v)
+
+    return output
+
+
+def _compute_sparse_attention_flash(
+    sparse_q: torch.Tensor,
+    k_blocks: torch.Tensor,
+    v_blocks: torch.Tensor,
+    kv_mask: torch.Tensor,
+) -> torch.Tensor:
+    """
+    FlashAttention implementation with per-batch, per-head mask support.
+
+    Strategy: check if all heads share the same mask. If so, use a single
+    FlashAttention call per batch (fast path). If not, process each head
+    separately (correct path).
+
+    Args:
+        sparse_q: [B, H, N, Sq, D]
+        k_blocks:  [B, H, N, Sk, D]
+        v_blocks:  [B, H, N, Sk, D]
+        kv_mask:   [B, H, N, N] boolean
+
+    Returns:
+        output: [B, H, N, Sq, D]
+    """
+    B, H, N, Sq, D = sparse_q.shape
+    Sk = k_blocks.shape[3]
+    device = sparse_q.device
+    output = torch.zeros_like(sparse_q)
+
+    for b in range(B):
+        # Check if all heads share the same mask for this batch element
+        # Compare each head's mask to head 0's mask
+        head0_mask = kv_mask[b, 0]  # [N, N]
+        all_heads_same = all(torch.equal(kv_mask[b, h], head0_mask) for h in range(1, H))
+
+        if all_heads_same:
+            # Fast path: all heads share the same mask, single FA call
+            _flash_attn_single_mask(
+                sparse_q[b],
+                k_blocks[b],
+                v_blocks[b],
+                head0_mask,
+                output[b],
+                H,
+                N,
+                Sq,
+                Sk,
+                D,
+                device,
+            )
+        else:
+            # Per-head path: process each head individually
+            for h in range(H):
+                head_mask = kv_mask[b, h]  # [N, N]
+                # Process single head: squeeze head dim, run FA, put back
+                _flash_attn_single_head(
+                    sparse_q[b, h],
+                    k_blocks[b, h],
+                    v_blocks[b, h],
+                    head_mask,
+                    output,
+                    b,
+                    h,
+                    N,
+                    Sq,
+                    Sk,
+                    D,
+                    device,
+                )
+
+    return output
+
+
+def _flash_attn_single_mask(
+    sparse_q_b: torch.Tensor,  # [H, N, Sq, D]
+    k_blocks_b: torch.Tensor,  # [H, N, Sk, D]
+    v_blocks_b: torch.Tensor,  # [H, N, Sk, D]
+    mask: torch.Tensor,  # [N, N] boolean
+    output_b: torch.Tensor,  # [H, N, Sq, D] (modified in-place)
+    H: int,
+    N: int,
+    Sq: int,
+    Sk: int,
+    D: int,
+    device: torch.device,
+) -> None:
+    """Run FlashAttention for all heads sharing the same KV mask."""
+    q_list = []
+    k_list = []
+    v_list = []
+    cu_seqlens_q = [0]
+    cu_seqlens_k = [0]
+    active_blocks = []
+
+    for qb in range(N):
+        selected = mask[qb]  # [N] boolean
+        sel_idx = selected.nonzero(as_tuple=True)[0]
+
+        if sel_idx.shape[0] == 0:
+            continue
+
+        active_blocks.append(qb)
+        num_kv_tokens = sel_idx.shape[0] * Sk
+
+        # [H, Sq, D] -> [Sq, H, D]
+        q_block = sparse_q_b[:, qb].permute(1, 0, 2)
+        q_list.append(q_block)
+
+        # [H, num_sel, Sk, D] -> [num_kv_tokens, H, D]
+        sel_k = k_blocks_b[:, sel_idx].permute(1, 2, 0, 3).reshape(num_kv_tokens, H, D)
+        sel_v = v_blocks_b[:, sel_idx].permute(1, 2, 0, 3).reshape(num_kv_tokens, H, D)
+        k_list.append(sel_k)
+        v_list.append(sel_v)
+
+        cu_seqlens_q.append(cu_seqlens_q[-1] + Sq)
+        cu_seqlens_k.append(cu_seqlens_k[-1] + num_kv_tokens)
+
+    if not q_list:
+        return
+
+    flat_q = torch.cat(q_list, dim=0)
+    flat_k = torch.cat(k_list, dim=0)
+    flat_v = torch.cat(v_list, dim=0)
+
+    cu_seqlens_q_t = torch.tensor(cu_seqlens_q, dtype=torch.int32, device=device)
+    cu_seqlens_k_t = torch.tensor(cu_seqlens_k, dtype=torch.int32, device=device)
+
+    max_seqlen_q = Sq
+    max_seqlen_k = int((cu_seqlens_k_t[1:] - cu_seqlens_k_t[:-1]).max().item())
+
+    orig_dtype = flat_q.dtype
+    compute_dtype = orig_dtype
+    if compute_dtype not in (torch.float16, torch.bfloat16):
+        compute_dtype = torch.bfloat16
+        flat_q = flat_q.to(compute_dtype)
+        flat_k = flat_k.to(compute_dtype)
+        flat_v = flat_v.to(compute_dtype)
+
+    flat_out = flash_attn_varlen_func_impl(
+        flat_q,
+        flat_k,
+        flat_v,
+        cu_seqlens_q_t,
+        cu_seqlens_k_t,
+        max_seqlen_q,
+        max_seqlen_k,
+        causal=False,
+    )
+
+    if compute_dtype != orig_dtype:
+        flat_out = flat_out.to(orig_dtype)
+
+    idx = 0
+    for qb in active_blocks:
+        block_out = flat_out[idx:idx + Sq]  # [Sq, H, D]
+        output_b[:, qb] = block_out.permute(1, 0, 2)  # [H, Sq, D]
+        idx += Sq
+
+
+def _flash_attn_single_head(
+    sparse_q_bh: torch.Tensor,  # [N, Sq, D]
+    k_blocks_bh: torch.Tensor,  # [N, Sk, D]
+    v_blocks_bh: torch.Tensor,  # [N, Sk, D]
+    mask: torch.Tensor,  # [N, N] boolean
+    output: torch.Tensor,  # [B, H, N, Sq, D] (modified in-place)
+    b: int,
+    h: int,
+    N: int,
+    Sq: int,
+    Sk: int,
+    D: int,
+    device: torch.device,
+) -> None:
+    """Run FlashAttention for a single head with its own KV mask."""
+    q_list = []
+    k_list = []
+    v_list = []
+    cu_seqlens_q = [0]
+    cu_seqlens_k = [0]
+    active_blocks = []
+
+    for qb in range(N):
+        selected = mask[qb]
+        sel_idx = selected.nonzero(as_tuple=True)[0]
+
+        if sel_idx.shape[0] == 0:
+            continue
+
+        active_blocks.append(qb)
+        num_kv_tokens = sel_idx.shape[0] * Sk
+
+        # [Sq, D] -> [Sq, 1, D] (single head)
+        q_block = sparse_q_bh[qb].unsqueeze(1)
+        q_list.append(q_block)
+
+        # [num_sel, Sk, D] -> [num_kv_tokens, 1, D]
+        sel_k = k_blocks_bh[sel_idx].reshape(num_kv_tokens, 1, D)
+        sel_v = v_blocks_bh[sel_idx].reshape(num_kv_tokens, 1, D)
+        k_list.append(sel_k)
+        v_list.append(sel_v)
+
+        cu_seqlens_q.append(cu_seqlens_q[-1] + Sq)
+        cu_seqlens_k.append(cu_seqlens_k[-1] + num_kv_tokens)
+
+    if not q_list:
+        return
+
+    flat_q = torch.cat(q_list, dim=0)
+    flat_k = torch.cat(k_list, dim=0)
+    flat_v = torch.cat(v_list, dim=0)
+
+    cu_seqlens_q_t = torch.tensor(cu_seqlens_q, dtype=torch.int32, device=device)
+    cu_seqlens_k_t = torch.tensor(cu_seqlens_k, dtype=torch.int32, device=device)
+
+    max_seqlen_q = Sq
+    max_seqlen_k = int((cu_seqlens_k_t[1:] - cu_seqlens_k_t[:-1]).max().item())
+
+    orig_dtype = flat_q.dtype
+    compute_dtype = orig_dtype
+    if compute_dtype not in (torch.float16, torch.bfloat16):
+        compute_dtype = torch.bfloat16
+        flat_q = flat_q.to(compute_dtype)
+        flat_k = flat_k.to(compute_dtype)
+        flat_v = flat_v.to(compute_dtype)
+
+    flat_out = flash_attn_varlen_func_impl(
+        flat_q,
+        flat_k,
+        flat_v,
+        cu_seqlens_q_t,
+        cu_seqlens_k_t,
+        max_seqlen_q,
+        max_seqlen_k,
+        causal=False,
+    )
+
+    if compute_dtype != orig_dtype:
+        flat_out = flat_out.to(orig_dtype)
+
+    idx = 0
+    for qb in active_blocks:
+        block_out = flat_out[idx:idx + Sq]  # [Sq, 1, D]
+        output[b, h, qb] = block_out.squeeze(1)  # [Sq, D]
+        idx += Sq
+
+
+def _reconstruct_pruned(
+    sparse_output: torch.Tensor,
+    keep_indices: torch.Tensor,
+    block_size: int,
+) -> torch.Tensor:
+    """
+    Scatter sparse output back to full block size.
+    Pruned positions get nearest kept token's output.
+
+    Handles per-batch, per-head indices correctly.
+
+    Args:
+        sparse_output: [B, H, N, keep_size, D]
+        keep_indices:  [B, H, N, keep_size]
+        block_size: original tokens per block
+
+    Returns:
+        full_output: [B, H, N, block_size, D]
+    """
+    B, H, N, keep_size, D = sparse_output.shape
+    device = sparse_output.device
+
+    if keep_size >= block_size:
+        return sparse_output
+
+    full_output = torch.zeros(B, H, N, block_size, D, device=device, dtype=sparse_output.dtype)
+
+    # Scatter kept tokens
+    idx_expand = keep_indices.unsqueeze(-1).expand(-1, -1, -1, -1, D)
+    full_output.scatter_(3, idx_expand, sparse_output)
+
+    # Fill pruned positions with nearest kept token (vectorized)
+    all_pos = torch.arange(block_size, device=device)
+
+    for b in range(B):
+        for h in range(H):
+            for n in range(N):
+                kept = keep_indices[b, h, n]  # [keep_size]
+
+                # Distance from every position to every kept position
+                dists = (all_pos.view(-1, 1) - kept.view(1, -1)).abs()
+                nearest_local_idx = dists.argmin(dim=1)  # [block_size]
+
+                # Identify pruned positions
+                is_pruned = torch.ones(block_size, dtype=torch.bool, device=device)
+                is_pruned[kept] = False
+                pruned_indices = is_pruned.nonzero(as_tuple=True)[0]
+
+                if pruned_indices.numel() > 0:
+                    src_indices = nearest_local_idx[pruned_indices]
+                    full_output[b, h, n, pruned_indices] = sparse_output[b, h, n, src_indices]
+
+    return full_output
+
+
+# ---------------------------------------------------------------------------
+# FastVideo backend classes
+# ---------------------------------------------------------------------------
+
+
+class BSAAttentionBackend(AttentionBackend):
+
+    accept_output_buffer: bool = False
+
+    @staticmethod
+    def get_supported_head_sizes() -> list[int]:
+        return [64, 128]
+
+    @staticmethod
+    def get_name() -> str:
+        return "BSA_ATTN"
+
+    @staticmethod
+    def get_impl_cls() -> type["BSAAttentionImpl"]:
+        return BSAAttentionImpl
+
+    @staticmethod
+    def get_metadata_cls() -> type["BSAAttentionMetadata"]:
+        return BSAAttentionMetadata
+
+    @staticmethod
+    def get_builder_cls() -> type["BSAAttentionMetadataBuilder"]:
+        return BSAAttentionMetadataBuilder
+
+
+@dataclass
+class BSAAttentionMetadata(AttentionMetadata):
+    current_timestep: int
+    dit_seq_shape: tuple[int, int, int]
+    total_seq_length: int
+    num_blocks: int
+    block_size: int
+    tile_partition_indices: torch.LongTensor
+    reverse_tile_partition_indices: torch.LongTensor
+    # BSA-specific config
+    query_keep_ratio: float
+    kv_cumulative_threshold: float
+    min_kv_blocks: int
+
+
+class BSAAttentionMetadataBuilder(AttentionMetadataBuilder):
+
+    def __init__(self):
+        pass
+
+    def prepare(self):
+        pass
+
+    def build(
+        self,
+        current_timestep: int,
+        raw_latent_shape: tuple[int, int, int],
+        patch_size: tuple[int, int, int],
+        device: torch.device,
+        bsa_query_keep_ratio: float = 0.5,
+        bsa_kv_cumulative_threshold: float = 0.9,
+        bsa_min_kv_blocks: int = 4,
+        **kwargs: dict[str, Any],
+    ) -> "BSAAttentionMetadata":
+        # Ensure patching does not drop tokens silently.
+        assert all(r % p == 0 for r, p in zip(raw_latent_shape, patch_size, strict=False)), (
+            "raw_latent_shape must be divisible by patch_size for BSA", )
+
+        dit_seq_shape = (
+            raw_latent_shape[0] // patch_size[0],
+            raw_latent_shape[1] // patch_size[1],
+            raw_latent_shape[2] // patch_size[2],
+        )
+
+        total_seq_length = math.prod(dit_seq_shape)
+        block_size = math.prod(BSA_TILE_SIZE)
+        # Require exact tiling to avoid reshape failures later.
+        assert all(d % t == 0 for d, t in zip(dit_seq_shape, BSA_TILE_SIZE, strict=False)), (
+            "dit_seq_shape must be divisible by BSA_TILE_SIZE", )
+        num_blocks = total_seq_length // block_size
+
+        tile_partition_indices = get_tile_partition_indices(dit_seq_shape, BSA_TILE_SIZE, device)
+        reverse_tile_partition_indices = get_reverse_tile_partition_indices(dit_seq_shape, BSA_TILE_SIZE, device)
+
+        return BSAAttentionMetadata(
+            current_timestep=current_timestep,
+            dit_seq_shape=dit_seq_shape,
+            total_seq_length=total_seq_length,
+            num_blocks=num_blocks,
+            block_size=block_size,
+            tile_partition_indices=tile_partition_indices,
+            reverse_tile_partition_indices=reverse_tile_partition_indices,
+            query_keep_ratio=bsa_query_keep_ratio,
+            kv_cumulative_threshold=bsa_kv_cumulative_threshold,
+            min_kv_blocks=bsa_min_kv_blocks,
+        )
+
+
+class BSAAttentionImpl(AttentionImpl):
+
+    def __init__(
+        self,
+        num_heads: int,
+        head_size: int,
+        causal: bool,
+        softmax_scale: float,
+        num_kv_heads: int | None = None,
+        prefix: str = "",
+        **extra_impl_args,
+    ) -> None:
+        self.prefix = prefix
+        self.num_heads = num_heads
+        self.head_size = head_size
+        if num_kv_heads is not None and num_kv_heads != num_heads:
+            raise ValueError("BSA backend does not support grouped-query attention")
+        if causal:
+            raise ValueError("BSA backend is bidirectional; causal=True is unsupported")
+        if softmax_scale is not None:
+            expected_scale = 1.0 / math.sqrt(self.head_size)
+            if not math.isclose(softmax_scale, expected_scale, rel_tol=1e-4, abs_tol=1e-5):
+                raise ValueError("softmax_scale must be default (1/sqrt(d)) for BSA")
+        try:
+            sp_group = get_sp_group()
+            self.sp_size = sp_group.world_size
+        except (AssertionError, RuntimeError):
+            self.sp_size = 1
+
+    def preprocess_qkv(
+        self,
+        qkv: torch.Tensor,
+        attn_metadata: BSAAttentionMetadata,
+    ) -> torch.Tensor:
+        """Reorder tokens from raster order to tile-contiguous order."""
+        # qkv: [B, L, num_heads, D]
+        return qkv[:, attn_metadata.tile_partition_indices]
+
+    def postprocess_output(
+        self,
+        output: torch.Tensor,
+        attn_metadata: BSAAttentionMetadata,
+    ) -> torch.Tensor:
+        """Reorder tokens from tile-contiguous order back to raster order."""
+        return output[:, attn_metadata.reverse_tile_partition_indices]
+
+    def forward(
+        self,
+        query: torch.Tensor,
+        key: torch.Tensor,
+        value: torch.Tensor,
+        attn_metadata: BSAAttentionMetadata,
+    ) -> torch.Tensor:
+        """
+        BSA attention forward pass.
+
+        Input tensors are already in tile-contiguous order from preprocess_qkv.
+
+        Args:
+            query: [B, L, num_heads, D] (tile-ordered)
+            key:   [B, L, num_heads, D] (tile-ordered)
+            value: [B, L, num_heads, D] (tile-ordered)
+            attn_metadata: BSA metadata
+
+        Returns:
+            output: [B, L, num_heads, D] (tile-ordered)
+        """
+        B, L, H, D = query.shape
+        block_size = attn_metadata.block_size
+        num_blocks = attn_metadata.num_blocks
+        assert num_blocks * block_size == L, "Sequence length must match tiling"
+
+        # Reshape to [B, H, L, D] for attention computation
+        q = query.transpose(1, 2).contiguous()  # [B, H, L, D]
+        k = key.transpose(1, 2).contiguous()
+        v = value.transpose(1, 2).contiguous()
+
+        # Reshape into blocks: [B, H, num_blocks, block_size, D]
+        q_blocks = q.view(B, H, num_blocks, block_size, D)
+        k_blocks = k.view(B, H, num_blocks, block_size, D)
+        v_blocks = v.view(B, H, num_blocks, block_size, D)
+
+        # --- Query sparsification ---
+        sparse_q, keep_indices, keep_size = _prune_queries(q_blocks, attn_metadata.query_keep_ratio)
+
+        # --- KV block selection ---
+        kv_mask = _select_kv_blocks(
+            sparse_q,
+            k_blocks,
+            attn_metadata.kv_cumulative_threshold,
+            attn_metadata.min_kv_blocks,
+        )
+
+        # --- Sparse attention ---
+        sparse_output = _compute_sparse_attention(sparse_q, k_blocks, v_blocks, kv_mask)
+
+        # --- Reconstruct pruned positions ---
+        full_output = _reconstruct_pruned(sparse_output, keep_indices, block_size)
+
+        # Reshape back: [B, H, num_blocks, block_size, D] -> [B, H, L, D] -> [B, L, H, D]
+        hidden_states = full_output.view(B, H, L, D).transpose(1, 2)
+
+        return hidden_states
@@ -72,12 +72,11 @@ class FlashAttnMetadataBuilder(AttentionMetadataBuilder):
        pass

    def build(  # type: ignore
-        self,
-        current_timestep: int,
-        attn_mask: torch.Tensor,
+            self,
+            current_timestep: int,
+            attn_mask: torch.Tensor,
    ) -> FlashAttnMetadata:
-        return FlashAttnMetadata(current_timestep=current_timestep,
-                                 attn_mask=attn_mask)
+        return FlashAttnMetadata(current_timestep=current_timestep, attn_mask=attn_mask)


 class FlashAttentionImpl(AttentionImpl):
@@ -103,32 +102,24 @@ class FlashAttentionImpl(AttentionImpl):
        attn_metadata: FlashAttnMetadata,
    ):

-        def _key_padding_mask_from_attn_mask(attn_mask: torch.Tensor,
-                                             key_len: int) -> torch.Tensor:
+        def _key_padding_mask_from_attn_mask(attn_mask: torch.Tensor, key_len: int) -> torch.Tensor:
            # Normalize attn_mask to [B, key_len] where True means valid token.
            if attn_mask.dim() == 4:
                attn_mask = attn_mask[:, 0, 0, :]
            elif attn_mask.dim() == 3:
                attn_mask = attn_mask[:, 0, :]
            elif attn_mask.dim() != 2:
-                raise ValueError(
-                    f"Unsupported attn_mask shape for FLASH_ATTN: {attn_mask.shape}"
-                )
+                raise ValueError(f"Unsupported attn_mask shape for FLASH_ATTN: {attn_mask.shape}")

-            if attn_mask.dtype == torch.bool:
-                key_padding_mask = attn_mask
-            else:
-                # SDPA additive mask convention: valid=0, masked=-inf/large negative.
-                key_padding_mask = attn_mask >= 0
+            # SDPA additive mask convention: valid=0, masked=-inf/large negative.
+            key_padding_mask = attn_mask if attn_mask.dtype == torch.bool else attn_mask >= 0

            if key_padding_mask.shape[-1] != key_len:
-                raise ValueError(
-                    "Invalid key padding mask length for FLASH_ATTN: "
-                    f"expected {key_len}, got {key_padding_mask.shape[-1]}")
+                raise ValueError("Invalid key padding mask length for FLASH_ATTN: "
+                                 f"expected {key_len}, got {key_padding_mask.shape[-1]}")
            return key_padding_mask

-        if (attn_metadata is not None and hasattr(attn_metadata, "attn_mask")
-                and attn_metadata.attn_mask is not None):
+        if (attn_metadata is not None and hasattr(attn_metadata, "attn_mask") and attn_metadata.attn_mask is not None):
            from fastvideo.attention.utils.flash_attn_no_pad import (
                flash_attn_no_pad,
                flash_attn_varlen_qk_no_pad,
@@ -144,8 +135,7 @@ class FlashAttentionImpl(AttentionImpl):
                    dtype=torch.bool,
                    device=query.device,
                )
-                key_padding_mask = _key_padding_mask_from_attn_mask(
-                    attn_mask, key.shape[1]).to(device=key.device)
+                key_padding_mask = _key_padding_mask_from_attn_mask(attn_mask, key.shape[1]).to(device=key.device)
                return flash_attn_varlen_qk_no_pad(
                    query,
                    key,
@@ -159,13 +149,8 @@ class FlashAttentionImpl(AttentionImpl):

            qkv = torch.stack([query, key, value], dim=2)

-            attn_mask = F.pad(attn_mask, (qkv.shape[1] - attn_mask.shape[1], 0),
-                              value=True)
-            output = flash_attn_no_pad(qkv,
-                                       attn_mask,
-                                       causal=False,
-                                       dropout_p=0,
-                                       softmax_scale=None)
+            attn_mask = F.pad(attn_mask, (qkv.shape[1] - attn_mask.shape[1], 0), value=True)
+            output = flash_attn_no_pad(qkv, attn_mask, causal=False, dropout_p=0, softmax_scale=None)
        else:
            output = flash_attn_func(
                query,  # type: ignore[no-untyped-call]
@@ -3,9 +3,7 @@
 import torch
 from sageattn3 import sageattn3_blackwell

-from fastvideo.attention.backends.abstract import (AttentionBackend,
-                                                   AttentionImpl,
-                                                   AttentionMetadata,
+from fastvideo.attention.backends.abstract import (AttentionBackend, AttentionImpl, AttentionMetadata,
                                                   AttentionMetadataBuilder)
 from fastvideo.logger import init_logger

@@ -3,8 +3,7 @@
 import torch
 from dataclasses import dataclass
 from fastvideo.attention.backends.abstract import (  # FlashAttentionMetadata,
-    AttentionBackend, AttentionImpl, AttentionMetadata,
-    AttentionMetadataBuilder)
+    AttentionBackend, AttentionImpl, AttentionMetadata, AttentionMetadataBuilder)
 from fastvideo.logger import init_logger

 logger = init_logger(__name__)
@@ -46,12 +45,11 @@ class SDPAMetadataBuilder(AttentionMetadataBuilder):
        pass

    def build(  # type: ignore
-        self,
-        current_timestep: int,
-        attn_mask: torch.Tensor,
+            self,
+            current_timestep: int,
+            attn_mask: torch.Tensor,
    ) -> SDPAMetadata:
-        return SDPAMetadata(current_timestep=current_timestep,
-                            attn_mask=attn_mask)
+        return SDPAMetadata(current_timestep=current_timestep, attn_mask=attn_mask)


 class SDPAImpl(AttentionImpl):
@@ -82,9 +80,8 @@ class SDPAImpl(AttentionImpl):
        key = key.transpose(1, 2)
        value = value.transpose(1, 2)

-        attn_mask = attn_metadata.attn_mask if (
-            attn_metadata is not None
-            and hasattr(attn_metadata, "attn_mask")) else None
+        attn_mask = attn_metadata.attn_mask if (attn_metadata is not None
+                                                and hasattr(attn_metadata, "attn_mask")) else None
        attn_kwargs = {
            "attn_mask": attn_mask,
            "dropout_p": self.dropout,
@@ -93,7 +90,6 @@ class SDPAImpl(AttentionImpl):
        }
        if query.shape[1] != key.shape[1]:
            attn_kwargs["enable_gqa"] = True
-        output = torch.nn.functional.scaled_dot_product_attention(
-            query, key, value, **attn_kwargs)
+        output = torch.nn.functional.scaled_dot_product_attention(query, key, value, **attn_kwargs)
        output = output.transpose(1, 2)
        return output
@@ -55,13 +55,11 @@ def compress_kernel(

    x_offset = idx_bh * L * D
    xm_offset = idx_bh * ((L + BLOCK_L - 1) // BLOCK_L) * D
-    x = tl.load(X + x_offset + offs_l[:, None] * D + offs_d[None, :],
-                mask=offs_l[:, None] < L)
+    x = tl.load(X + x_offset + offs_l[:, None] * D + offs_d[None, :], mask=offs_l[:, None] < L)

    nx = min(BLOCK_L, L - idx_l * BLOCK_L)
    x_mean = tl.sum(x, axis=0, dtype=tl.float32) / nx
-    tl.store(XM + xm_offset + idx_l * D + offs_d,
-             x_mean.to(XM.dtype.element_ty))
+    tl.store(XM + xm_offset + idx_l * D + offs_d, x_mean.to(XM.dtype.element_ty))


 def mean_pool(x: torch.Tensor, BLK: int) -> torch.Tensor:
@@ -98,8 +96,7 @@ def get_block_map(
        lut: Top-k indices of shape (B, H, num_q_blocks, topk)
        topk: Number of key blocks selected
    """
-    arg_k = k - torch.mean(
-        k, dim=-2, keepdim=True)  # smooth-k technique from SageAttention
+    arg_k = k - torch.mean(k, dim=-2, keepdim=True)  # smooth-k technique from SageAttention
    pooled_qblocks = mean_pool(q, BLKQ)
    pooled_kblocks = mean_pool(arg_k, BLKK)
    pooled_score = pooled_qblocks @ pooled_kblocks.transpose(-1, -2)
@@ -300,11 +297,7 @@ class SLAAttentionImpl(AttentionImpl, nn.Module):
            topk_ratio = attn_metadata.topk_ratio  # type: ignore[union-attr]

        # Compute block-sparse attention pattern
-        sparse_map, lut, real_topk = get_block_map(q,
-                                                   k,
-                                                   topk_ratio=topk_ratio,
-                                                   BLKQ=self.BLKQ,
-                                                   BLKK=self.BLKK)
+        sparse_map, lut, real_topk = get_block_map(q, k, topk_ratio=topk_ratio, BLKQ=self.BLKQ, BLKK=self.BLKK)

        # Convert to compute dtype
        q = q.to(self.dtype)
@@ -312,8 +305,7 @@ class SLAAttentionImpl(AttentionImpl, nn.Module):
        v = v.to(self.dtype)

        # Sparse attention
-        o_s = _attention.apply(q, k, v, sparse_map, lut, real_topk, self.BLKQ,
-                               self.BLKK)
+        o_s = _attention.apply(q, k, v, sparse_map, lut, real_topk, self.BLKQ, self.BLKK)

        # Linear attention with feature maps
        q_linear = self.feature_map_q(q).contiguous().to(self.dtype)
@@ -412,14 +404,10 @@ class SageSLAAttentionImpl(AttentionImpl, nn.Module):
        nn.Module.__init__(self)

        if not SAGESLA_ENABLED:
-            raise ImportError(
-                "SageSLA requires spas_sage_attn. "
-                "Install with: pip install git+https://github.com/thu-ml/SpargeAttn.git"
-            )
+            raise ImportError("SageSLA requires spas_sage_attn. "
+                              "Install with: pip install git+https://github.com/thu-ml/SpargeAttn.git")

-        assert head_size in [
-            64, 128
-        ], f"SageSLA requires head_size in [64, 128], got {head_size}"
+        assert head_size in [64, 128], f"SageSLA requires head_size in [64, 128], got {head_size}"

        self.num_heads = num_heads
        self.head_size = head_size
@@ -509,11 +497,7 @@ class SageSLAAttentionImpl(AttentionImpl, nn.Module):
            BLKQ, BLKK = 128, 64

        # Compute block-sparse attention pattern
-        sparse_map, lut, real_topk = get_block_map(q,
-                                                   k,
-                                                   topk_ratio=topk_ratio,
-                                                   BLKQ=BLKQ,
-                                                   BLKK=BLKK)
+        sparse_map, lut, real_topk = get_block_map(q, k, topk_ratio=topk_ratio, BLKQ=BLKQ, BLKK=BLKK)

        # Convert to compute dtype
        q = q.to(self.dtype)
@@ -526,47 +510,33 @@ class SageSLAAttentionImpl(AttentionImpl, nn.Module):
        scale = 1.0 / (headdim**0.5)

        # Quantize Q, K to INT8
-        q_int8, q_scale, k_int8, k_scale = get_vanilla_qk_quant(
-            q, k, km, BLKQ, BLKK)
+        q_int8, q_scale, k_int8, k_scale = get_vanilla_qk_quant(q, k, km, BLKQ, BLKK)
        lut_triton, valid_block_num = block_map_lut_triton(sparse_map)

        # Quantize V to FP8
        b, h_kv, kv_len, head_dim = v.shape
        padded_len = (kv_len + 127) // 128 * 128
-        v_transposed_permutted = torch.empty((b, h_kv, head_dim, padded_len),
-                                             dtype=v.dtype,
-                                             device=v.device)
+        v_transposed_permutted = torch.empty((b, h_kv, head_dim, padded_len), dtype=v.dtype, device=v.device)
        fused.transpose_pad_permute_cuda(v, v_transposed_permutted, 1)
-        v_fp8 = torch.empty(v_transposed_permutted.shape,
-                            dtype=torch.float8_e4m3fn,
-                            device=v.device)
-        v_scale = torch.empty((b, h_kv, head_dim),
-                              dtype=torch.float32,
-                              device=v.device)
-        fused.scale_fuse_quant_cuda(v_transposed_permutted, v_fp8, v_scale,
-                                    kv_len, 2.25, 1)
+        v_fp8 = torch.empty(v_transposed_permutted.shape, dtype=torch.float8_e4m3fn, device=v.device)
+        v_scale = torch.empty((b, h_kv, head_dim), dtype=torch.float32, device=v.device)
+        fused.scale_fuse_quant_cuda(v_transposed_permutted, v_fp8, v_scale, kv_len, 2.25, 1)

        # Sparse attention with quantized kernels
        o_s = torch.empty_like(q)
        if arch == "sm90":
            qattn.qk_int8_sv_f8_accum_f32_block_sparse_attn_inst_buf_fuse_v_scale_sm90(
-                q_int8, k_int8, v_fp8, o_s, lut_triton, valid_block_num,
-                q_scale, k_scale, v_scale, 1, False, 1, scale)
+                q_int8, k_int8, v_fp8, o_s, lut_triton, valid_block_num, q_scale, k_scale, v_scale, 1, False, 1, scale)
        else:
-            pvthreshold = torch.full((q.shape[-3], ),
-                                     1e6,
-                                     dtype=torch.float32,
-                                     device=q.device)
+            pvthreshold = torch.full((q.shape[-3], ), 1e6, dtype=torch.float32, device=q.device)
            if SAGE2PP_ENABLED:
                qk_int8_sv_f8_accum_f16_block_sparse_attn_inst_buf_fuse_v_scale_with_pv_threshold(
-                    q_int8, k_int8, v_fp8, o_s, lut_triton, valid_block_num,
-                    pvthreshold, q_scale, k_scale, v_scale, 1, False, 1, scale,
-                    0)
+                    q_int8, k_int8, v_fp8, o_s, lut_triton, valid_block_num, pvthreshold, q_scale, k_scale, v_scale, 1,
+                    False, 1, scale, 0)
            else:
                qattn.qk_int8_sv_f8_accum_f32_block_sparse_attn_inst_buf_fuse_v_scale_with_pv_threshold(
-                    q_int8, k_int8, v_fp8, o_s, lut_triton, valid_block_num,
-                    pvthreshold, q_scale, k_scale, v_scale, 1, False, 1, scale,
-                    0)
+                    q_int8, k_int8, v_fp8, o_s, lut_triton, valid_block_num, pvthreshold, q_scale, k_scale, v_scale, 1,
+                    False, 1, scale, 0)
        # ========== END SPARGE ==========

        # Linear attention with feature maps
@@ -12,9 +12,7 @@ except ImportError:

 from typing import Any

-from fastvideo.attention.backends.abstract import (AttentionBackend,
-                                                   AttentionImpl,
-                                                   AttentionMetadata,
+from fastvideo.attention.backends.abstract import (AttentionBackend, AttentionImpl, AttentionMetadata,
                                                   AttentionMetadataBuilder)
 from fastvideo.distributed import get_sp_group
 from fastvideo.logger import init_logger
@@ -31,14 +29,12 @@ def get_tile_partition_indices(
 ) -> torch.LongTensor:
    T, H, W = dit_seq_shape
    ts, hs, ws = tile_size
-    indices = torch.arange(T * H * W, device=device,
-                           dtype=torch.long).reshape(T, H, W)
+    indices = torch.arange(T * H * W, device=device, dtype=torch.long).reshape(T, H, W)
    ls = []
    for t in range(math.ceil(T / ts)):
        for h in range(math.ceil(H / hs)):
            for w in range(math.ceil(W / ws)):
-                ls.append(indices[t * ts:min(t * ts + ts, T),
-                                  h * hs:min(h * hs + hs, H),
+                ls.append(indices[t * ts:min(t * ts + ts, T), h * hs:min(h * hs + hs, H),
                                  w * ws:min(w * ws + ws, W)].flatten())
    index = torch.cat(ls, dim=0)
    return index
@@ -50,8 +46,7 @@ def get_reverse_tile_partition_indices(
    tile_size: tuple[int, int, int],
    device: torch.device,
 ) -> torch.LongTensor:
-    return torch.argsort(
-        get_tile_partition_indices(dit_seq_shape, tile_size, device))
+    return torch.argsort(get_tile_partition_indices(dit_seq_shape, tile_size, device))


@functools.lru_cache(maxsize=10)
@@ -104,10 +99,8 @@ def get_non_pad_index(
    n_win = variable_block_sizes.shape[0]
    device = variable_block_sizes.device
    starts_pad = torch.arange(n_win, device=device) * max_block_size
-    index_pad = starts_pad[:, None] + torch.arange(max_block_size,
-                                                   device=device)[None, :]
-    index_mask = torch.arange(
-        max_block_size, device=device)[None, :] < variable_block_sizes[:, None]
+    index_pad = starts_pad[:, None] + torch.arange(max_block_size, device=device)[None, :]
+    index_mask = torch.arange(max_block_size, device=device)[None, :] < variable_block_sizes[:, None]
    return index_pad[index_mask]


@@ -167,23 +160,17 @@ class VideoSparseAttentionMetadataBuilder(AttentionMetadataBuilder):
        **kwargs: dict[str, Any],
    ) -> VideoSparseAttentionMetadata:
        patch_size = patch_size
-        dit_seq_shape = (raw_latent_shape[0] // patch_size[0],
-                         raw_latent_shape[1] // patch_size[1],
+        dit_seq_shape = (raw_latent_shape[0] // patch_size[0], raw_latent_shape[1] // patch_size[1],
                         raw_latent_shape[2] // patch_size[2])

-        num_tiles = (math.ceil(dit_seq_shape[0] / VSA_TILE_SIZE[0]),
-                     math.ceil(dit_seq_shape[1] / VSA_TILE_SIZE[1]),
+        num_tiles = (math.ceil(dit_seq_shape[0] / VSA_TILE_SIZE[0]), math.ceil(dit_seq_shape[1] / VSA_TILE_SIZE[1]),
                     math.ceil(dit_seq_shape[2] / VSA_TILE_SIZE[2]))
        total_seq_length = math.prod(dit_seq_shape)

-        tile_partition_indices = get_tile_partition_indices(
-            dit_seq_shape, VSA_TILE_SIZE, device)
-        reverse_tile_partition_indices = get_reverse_tile_partition_indices(
-            dit_seq_shape, VSA_TILE_SIZE, device)
-        variable_block_sizes = construct_variable_block_sizes(
-            dit_seq_shape, num_tiles, device)
-        non_pad_index = get_non_pad_index(variable_block_sizes,
-                                          math.prod(VSA_TILE_SIZE))
+        tile_partition_indices = get_tile_partition_indices(dit_seq_shape, VSA_TILE_SIZE, device)
+        reverse_tile_partition_indices = get_reverse_tile_partition_indices(dit_seq_shape, VSA_TILE_SIZE, device)
+        variable_block_sizes = construct_variable_block_sizes(dit_seq_shape, num_tiles, device)
+        non_pad_index = get_non_pad_index(variable_block_sizes, math.prod(VSA_TILE_SIZE))

        return VideoSparseAttentionMetadata(
            current_timestep=current_timestep,
@@ -213,23 +200,19 @@ class VideoSparseAttentionImpl(AttentionImpl):
        sp_group = get_sp_group()
        self.sp_size = sp_group.world_size

-    def tile(self, x: torch.Tensor, num_tiles: list[int],
-             tile_partition_indices: torch.LongTensor,
+    def tile(self, x: torch.Tensor, num_tiles: list[int], tile_partition_indices: torch.LongTensor,
             non_pad_index: torch.LongTensor) -> torch.Tensor:
        t_padded_size = num_tiles[0] * VSA_TILE_SIZE[0]
        h_padded_size = num_tiles[1] * VSA_TILE_SIZE[1]
        w_padded_size = num_tiles[2] * VSA_TILE_SIZE[2]

-        x_padded = torch.zeros(
-            (x.shape[0], t_padded_size * h_padded_size * w_padded_size,
-             x.shape[-2], x.shape[-1]),
-            device=x.device,
-            dtype=x.dtype)
+        x_padded = torch.zeros((x.shape[0], t_padded_size * h_padded_size * w_padded_size, x.shape[-2], x.shape[-1]),
+                               device=x.device,
+                               dtype=x.dtype)
        x_padded[:, non_pad_index] = x[:, tile_partition_indices]
        return x_padded

-    def untile(self, x: torch.Tensor,
-               reverse_tile_partition_indices: torch.LongTensor,
+    def untile(self, x: torch.Tensor, reverse_tile_partition_indices: torch.LongTensor,
               non_pad_index: torch.LongTensor) -> torch.Tensor:
        x = x[:, non_pad_index][:, reverse_tile_partition_indices]
        return x
@@ -239,8 +222,7 @@ class VideoSparseAttentionImpl(AttentionImpl):
        qkv: torch.Tensor,
        attn_metadata: VideoSparseAttentionMetadata,
    ) -> torch.Tensor:
-        return self.tile(qkv, attn_metadata.num_tiles,
-                         attn_metadata.tile_partition_indices,
+        return self.tile(qkv, attn_metadata.num_tiles, attn_metadata.tile_partition_indices,
                         attn_metadata.non_pad_index)

    def postprocess_output(
@@ -248,8 +230,7 @@ class VideoSparseAttentionImpl(AttentionImpl):
        output: torch.Tensor,
        attn_metadata: VideoSparseAttentionMetadata,
    ) -> torch.Tensor:
-        return self.untile(output, attn_metadata.reverse_tile_partition_indices,
-                           attn_metadata.non_pad_index)
+        return self.untile(output, attn_metadata.reverse_tile_partition_indices, attn_metadata.non_pad_index)

    def forward(  # type: ignore[override]
        self,
@@ -266,20 +247,17 @@ class VideoSparseAttentionImpl(AttentionImpl):

        VSA_sparsity = attn_metadata.VSA_sparsity

-        cur_topk = math.ceil(
-            (1 - VSA_sparsity) *
-            (attn_metadata.total_seq_length / math.prod(VSA_TILE_SIZE)))
+        cur_topk = math.ceil((1 - VSA_sparsity) * (attn_metadata.total_seq_length / math.prod(VSA_TILE_SIZE)))

        if video_sparse_attn is None:
            raise NotImplementedError("video_sparse_attn is not installed")
-        hidden_states = video_sparse_attn(
-            query,
-            key,
-            value,
-            attn_metadata.variable_block_sizes,
-            attn_metadata.variable_block_sizes,
-            cur_topk,
-            block_size=VSA_TILE_SIZE,
-            compress_attn_weight=gate_compress).transpose(1, 2)
+        hidden_states = video_sparse_attn(query,
+                                          key,
+                                          value,
+                                          attn_metadata.variable_block_sizes,
+                                          attn_metadata.variable_block_sizes,
+                                          cur_topk,
+                                          block_size=VSA_TILE_SIZE,
+                                          compress_attn_weight=gate_compress).transpose(1, 2)

        return hidden_states
@@ -6,11 +6,8 @@ from dataclasses import dataclass
 import torch
 from einops import rearrange

-from fastvideo_kernel import (moba_attn_varlen, process_moba_input,
-                              process_moba_output)
-from fastvideo.attention.backends.abstract import (AttentionBackend,
-                                                   AttentionImpl,
-                                                   AttentionMetadata,
+from fastvideo_kernel import (moba_attn_varlen, process_moba_input, process_moba_output)
+from fastvideo.attention.backends.abstract import (AttentionBackend, AttentionImpl, AttentionMetadata,
                                                   AttentionMetadataBuilder)
 from fastvideo.logger import init_logger

@@ -94,12 +91,10 @@ class VideoMobaAttentionMetadataBuilder(AttentionMetadataBuilder):
    ) -> VideoMobaAttentionMetadata:
        if device is None:
            device = torch.device("cpu")
-        assert raw_latent_shape[0] % patch_size[0] == 0 and raw_latent_shape[
-            1] % patch_size[1] == 0 and raw_latent_shape[2] % patch_size[
+        assert raw_latent_shape[0] % patch_size[0] == 0 and raw_latent_shape[1] % patch_size[
+            1] == 0 and raw_latent_shape[2] % patch_size[
                2] == 0, f"spatial patch_resolution {raw_latent_shape} should be divisible by patch_size {patch_size}"
-        patch_resolution = [
-            t // pt for t, pt in zip(raw_latent_shape, patch_size, strict=False)
-        ]
+        patch_resolution = [t // pt for t, pt in zip(raw_latent_shape, patch_size, strict=False)]

        return VideoMobaAttentionMetadata(
            current_timestep=current_timestep,
@@ -170,19 +165,11 @@ class VMOBAAttentionImpl(AttentionImpl):
            moba_chunk_size = attn_metadata.st_chunk_size
            moba_topk = attn_metadata.st_topk

-        query, chunk_size = process_moba_input(query,
-                                               attn_metadata.patch_resolution,
-                                               moba_chunk_size)
-        key, chunk_size = process_moba_input(key,
-                                             attn_metadata.patch_resolution,
-                                             moba_chunk_size)
-        value, chunk_size = process_moba_input(value,
-                                               attn_metadata.patch_resolution,
-                                               moba_chunk_size)
+        query, chunk_size = process_moba_input(query, attn_metadata.patch_resolution, moba_chunk_size)
+        key, chunk_size = process_moba_input(key, attn_metadata.patch_resolution, moba_chunk_size)
+        value, chunk_size = process_moba_input(value, attn_metadata.patch_resolution, moba_chunk_size)
        max_seqlen = query.shape[1]
-        indices_q = torch.arange(0,
-                                 query.shape[0] * query.shape[1],
-                                 device=query.device)
+        indices_q = torch.arange(0, query.shape[0] * query.shape[1], device=query.device)
        cu_seqlens = torch.arange(0,
                                  query.shape[0] * query.shape[1] + 1,
                                  query.shape[1],
@@ -205,10 +192,7 @@ class VMOBAAttentionImpl(AttentionImpl):
            simsum_threshold=attn_metadata.moba_threshold,
            threshold_type=attn_metadata.moba_threshold_type,
        )
-        hidden_states = self.pad_input(hidden_states, indices_q, batch_size,
-                                       sequence_length)
-        hidden_states = process_moba_output(hidden_states,
-                                            attn_metadata.patch_resolution,
-                                            moba_chunk_size)
+        hidden_states = self.pad_input(hidden_states, indices_q, batch_size, sequence_length)
+        hidden_states = process_moba_output(hidden_states, attn_metadata.patch_resolution, moba_chunk_size)

        return hidden_states
--- a/Show More
+++ b/Show More
Author	SHA1	Message	Date
SolitaryThinkerandClaude Opus 4.6	300eab2b8d	[feat] Migrate Hunyuan SamplingParam subclasses to profile-based defaults Replace HunyuanSamplingParam and FastHunyuanSamplingParam subclasses with profile-based defaults. This adds the profile infrastructure (ConfigInfo.default_profile, SamplingParam._from_profile) and the first profile definitions for Hunyuan models, enabling future model families to migrate from subclass overrides to declarative profiles. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>	2026-04-15 16:33:37 -07:00
William Lin	88a5a933ab	[feat] [3/n] Improve API: extend support to cli (#1226 )	2026-04-14 15:20:47 -07:00
William Lin	c591d6d2a6	[feat] [2/n] Improve API: add initial support in video_generator (#1220 )	2026-04-06 10:33:54 -07:00
Kun Linandmergify[bot]	65dff806a8	[bugfix]Fixing Lora distillation training distributed checkpointing bug (#1192 ) Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>	2026-04-06 02:20:26 +00:00
KUAN-HAO HUANGandmergify[bot]	b85f0f4c2a	[perf]: Eliminate CPU-GPU synchronization bottlenecks in training pipeline (#1217 ) Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>	2026-04-06 02:03:46 +00:00
William Lin	76c62d7a00	[feat] [1/n] API improvements: add intial files for new fastvideo public API (#1218 )	2026-04-05 18:13:19 -07:00
Satyam Srivastava Satyam Srivastava Satyam Srivastava mergify[bot]	f6e65ff668	[Feature] Add BSA (Bidirectional Sparse Attention) inference backend (#1174 ) Co-authored-by: Satyam Srivastava <satyam53@Mac.lan1> Co-authored-by: Satyam Srivastava <satyam53@Satyams-MacBook-Air.local> Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>	2026-04-05 05:00:33 +00:00
mergify[bot]	c220aa8000	[ci](mergify): upgrade configuration to current format (#1216 ) Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>	2026-04-04 23:09:17 +00:00
Jinzhe PanandDarren Sadr	4713fc17ed	[feat] Job Runner UI (#1189 ) Co-authored-by: Darren Sadr <darrensadr@gmail.com>	2026-04-02 16:07:24 -07:00
vishruthb	5789955bbe	[feat] add gen3c (cosmos-7b) model and pipeline support (#1059 )	2026-04-01 11:42:02 +00:00
Jinzhe Pan	2ad84a3b78	[ci] Use update instead of rebase for auto branch sync (#1215 )	2026-04-01 19:16:59 +08:00
Jinzhe Pan	12d699cd78	[ci] Add direct test retry with check overwrite and aggregate status refresh (#1214 )	2026-04-01 17:21:28 +08:00
Jinzhe Pan	34f14ded21	[ci] Use pull_request_target for Full Suite trigger (#1213 )	2026-04-01 03:01:07 +08:00
Jinzhe Pan	71d1ab411f	[ci] Fix jq crash when Buildkite build env is null (#1212 )	2026-04-01 02:35:01 +08:00
Jinzhe Pan	805e487773	[ci] Ignore legacy reference videos when checking for HF download (#1211 )	2026-04-01 02:12:09 +08:00
Jinzhe Pan	8803b4547e	[ci] Add retry for flaky tests and fix stale SSIM references (#1210 )	2026-04-01 01:11:49 +08:00
Jinzhe Pan	3b3806b3f6	[ci] Fix /merge to directly trigger Full Suite + simplify rebase conditions (#1209 )	2026-03-31 23:17:09 +08:00
Jinzhe Pan	38d962e89d	[ci] Remove Mergify ready-label race condition (#1208 )	2026-03-31 20:59:13 +08:00
Jinzhe Pan	3966a365d0	[ci] Add statuses:write permission for /test pre-commit (#1207 )	2026-03-31 20:33:18 +08:00
Jinzhe Pan	d73fd14af0	[ci] Post pre-commit status to PR commit SHA (#1206 )	2026-03-31 20:21:21 +08:00
Jinzhe Pan	a87cc89916	[ci] Trigger pre-commit on /test slash commands (#1205 )	2026-03-31 20:12:57 +08:00
Jinzhe Pan	81fd80c8ee	[ci] Add TEST_SCOPE routing for clean single-test execution (#1203 )	2026-03-31 19:40:59 +08:00
Jinzhe Pan	ff22439f28	[ci] Fix fork PR checkout for /test and Full Suite triggers (#1202 )	2026-03-31 13:42:13 +08:00
Jinzhe Pan	de0de04212	[ci] Replace Merge Queue with auto-merge — reduce CI complexity (#1200 )	2026-03-31 10:09:33 +08:00
mergify[bot]	7f2c3e1f64	[ci](mergify): upgrade configuration to current format (#1194 ) Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>	2026-03-31 00:47:48 +08:00
Jinzhe Pan	ab55e57c22	[ci] Fix Merge Queue requeue and draft PR pre-commit skip (#1197 )	2026-03-30 22:35:13 +08:00
Jinzhe Pan	46f6b43a53	[ci] Fix Merge Queue immediate dequeue (#1196 )	2026-03-30 21:33:45 +08:00
Jinzhe Pan	833a33b663	[ci] CI follow-up: gate checks, issue label unification, draft PR skip (#1193 )	2026-03-30 20:19:46 +08:00
Jinzhe Pan	9ea1307cd4	[ci] Add approval and pre-commit checks to merge protections (#1190 ) ## Summary Follow-up to #1187. Two small changes: 1. Merge Protections expanded — adds `#approved-reviews-by>=1` and `check-success~=pre-commit` to `merge_protections` so the Mergify check shows a unified requirements checklist on every PR (title format + approval + pre-commit), instead of only showing the title format. 2. Buildkite pipeline comment fix — updates the outdated Full Suite section comment from "Triggered by adding the 'ready' label via GitHub Actions → Buildkite API" to reflect the new Merge Queue trigger path.	2026-03-30 05:14:49 +00:00
Jinzhe Pan	be35003cb1	[ci] Merge Queue, label system overhaul, and slash commands (2/2) (#1187 )	2026-03-30 08:22:35 +08:00
Jinzhe Pan	26bd4db253	[ci] CI infrastructure cleanup and workflow reorganization (1/2) (#1186 )	2026-03-29 17:01:09 -07:00
Jinzhe PanandWill Lin	e294ca011c	[feat]: overhaul SSIM test infrastructure — partition scheduling, helper migration, CI fixes (#1185 ) Co-authored-by: Will Lin <wlsaidhi@gmail.com>	2026-03-29 23:59:25 +00:00
alexzms	2085a4fc4a	[bugfix]: fix VAE temporal tiling blend corruption in tiled_encode (#1181 )	2026-03-29 23:12:42 +00:00
Jinzhe Pan	c0c8e39c04	Revert "[feat] Job Runner UI" (#1188 )	2026-03-29 16:46:27 +08:00
Darren	f72618dafb	[feat] Job Runner UI (#1172 )	2026-03-29 16:17:56 +08:00
alexzms	b3edfacdd8	[bugfix]: fix I2V preprocessing crash for models without CLIP (Wan2.2 I2V) (#1184 )	2026-03-28 10:28:50 +08:00
alexzms	30129a3350	[misc]: reorganize training configs and add documentation (#1177 )	2026-03-26 16:38:31 -07:00
jaisurya27	4d49f7b0aa	Kandinsky5 lite dit clean (#1088 )	2026-03-26 08:07:11 +08:00
alexzms	71bfc13d75	[feat]: add HunyuanVideo model plugin for fastvideo/train framework (#1175 )	2026-03-24 16:49:51 -07:00
Kaiqin Kong	74db6e18d1	[misc] update action loading in validation and preprocess (#1143 )	2026-03-24 15:10:03 -07:00
Kaiqin Kong	7d263c6a36	[bugfix] self-forcing train/validation step mismatch (#1173 )	2026-03-20 00:52:45 -07:00
Zhang Peiyuan	454c32d1d1	Update README.md	2026-03-17 14:26:07 -07:00
Jinzhe Pan	d1240b9238	[CI] add contributor interaction automation (#1170 )	2026-03-17 12:05:40 +08:00
Hao Zhang	4105094fa5	[docs] Update README with realtime demo announcement (#1169 )	2026-03-13 15:52:02 -07:00
alexzms	f036469d3d	[feat]: Knowledge Distillation training method for ODE-init (KDMethod + KDCausalMethod) (#1166 )	2026-03-11 20:43:24 -07:00
alexzms	14261bc98c	[feat] pre-commit support 120 col num (#1167 )	2026-03-11 20:19:30 -07:00
				`@@ -0,0 +1 @@`
				`WRN 2026-03-26T13:46:33.469 ?.19646 server_start:193: Failed to start server: operation not permitted: /var/folders/z_/h_6myyk14d1b7z87z3vy4mjh0000gn/T/nvim.dsynkd/iSe0el/nvim.19646.0`