generation
Route waveform diffusion generator, text-conditioning, sampling, inpainting, and DiffusionAR tasks.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- generation
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Route waveform diffusion generator, text-conditioning, sampling, inpainting, and DiffusionAR tasks.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/api-reference.md4.1 KB
- 参考文档references/troubleshooting.md3.2 KB
- 参考文档references/workflows.md3.2 KB
- 脚本scripts/tiny_generation_smoke.py4.5 KB
03 SKILL.md 正文
Generation
Use this sub-skill when a request is about:
- building or smoke-checking an unconditional audio diffusion generator,
- sampling with
DiffusionModel.sample, - enabling text conditioning with classifier-free guidance,
- filling masked waveform gaps with
VInpainter, - or handling the exported but expert-only
DiffusionARpath.
Route elsewhere:
- upsampling, vocoding, and autoencoding belong in
../conditioning/. - pretrained checkpoints, full training runs, benchmark sweeps, and paper-specific recipes are not part of this sub-skill.
- release and import workflows belong to the repo-level skill, not this route.
Quick workflow
- Start with
scripts/tiny_generation_smoke.pyfor a CPU-only, no-pretrained-weights check. - Use
references/api-reference.mdto confirm constructor names, argument rules, and route behavior. - Use
references/workflows.mdfor the smallest working shapes and expected signals. - Use
references/troubleshooting.mdwhen you hit channel, mask, text-conditioning, or step/schedule errors.
Owned capabilities
DiffusionModel.forwardloss route andDiffusionModel.samplesampler route.UNetV0construction, length checks, attention and cross-attention flags, text-conditioning flags, andresnet_groupsconstraints.VDiffusion,UniformDistribution,LinearSchedule, andVSampler.VInpaintersource/mask/noise resampling flow.- Expert notes for
DiffusionAR.
Bundled files
references/api-reference.mdreferences/workflows.mdreferences/troubleshooting.mdscripts/tiny_generation_smoke.py
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).