video-and-audio-apps
Route PaddleGAN video, motion, and lip-sync workflows.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- video-and-audio-apps
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Route PaddleGAN video, motion, and lip-sync workflows.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/motion-and-lipsync.md3.0 KB
- 参考文档references/troubleshooting.md2.4 KB
- 参考文档references/video-workflows.md2.9 KB
- 脚本scripts/check_video_stack.py7.6 KB
03 SKILL.md 正文
video-and-audio-apps
Use this sub-skill for PaddleGAN workflows that transform, restore, interpolate, or synchronize video and audio.
Owns
- video colorization and restoration
- video frame interpolation
- video super-resolution
- First Order Motion
- Wav2Lip lip-sync
- composite planning across those workflows
Excludes
- single-image-only app requests -> image-and-face-apps
- exported static-model inference or export -> deployment-export
- dataset preparation for LRS2, REDS, Vimeo90K, or similar corpora -> data-preparation
First move
- Run
scripts/check_video_stack.pyto confirm ffmpeg, imageio, librosa, Paddle, and face-detector readiness. - Read
references/video-workflows.mdfor model-family routing and composite planning. - Read
references/motion-and-lipsync.mdfor First Order Motion and Wav2Lip flags. - Use
references/troubleshooting.mdbefore retrying a failure.
Runtime rules
- Use public
ppgan.appspredictors or bundled helpers; treat unbundled demo scripts as reference-only planning material. - Do not assume automatic weight downloads, GPU memory headroom, or ffmpeg availability.
- Defer heavy execution until stack checks pass and the clip, weights, and device budget are ready.
- Prefer CPU only for readiness checks; use GPU for real media runs when available.
Core predictor families
DeOldifyPredictorDeepRemasterPredictorDAINPredictorRealSRPredictorEDVRPredictorBasicVSRPredictorIconVSRPredictorBasiVSRPlusPlusPredictorPPMSVSRPredictorPPMSVSRLargePredictorFirstOrderPredictorWav2LipPredictor
Planning note
The legacy composite order vocabulary is DAIN, DeepRemaster, DeOldify, RealSR, EDVR, BasicVSR, IconVSR, BasiVSRPlusPlus, PPMSVSR, and PPMSVSRLarge; plan with those names, but rely on the bundled guidance rather than the unbundled demo CLI.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).