[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M796RHVCXK5GQYM0N63MNVWC":3,"skill-same:01M796RHVCXK5GQYM0N63MNVWC":75},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":33,"install":34,"resources":46},"01M796RHVCXK5GQYM0N63MNVWC","turbo-diffusion",null,"Use TurboDiffusion for accelerated Wan video generation, interactive serving, checkpoint conversion, CUDA acceleration backends, and TurboT2AV extension planning.","VectorSpaceLab\u002FAREX-Skill","VectorSpaceLab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F180637102?v=4","01M7538C5WPSJGWF7KHR4EFZ9F","01M7538C6R2GFP53NPXTCZNBRT",351,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: turbo-diffusion\ndescription: \"Use TurboDiffusion for accelerated Wan video generation,\n  interactive serving, checkpoint conversion, CUDA acceleration backends, and\n  TurboT2AV extension planning.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# TurboDiffusion\n\nUse this repo skill when a task involves TurboDiffusion, TurboWan, Wan2.1\u002FWan2.2 accelerated video generation, SLA\u002FSageSLA attention, rCM\u002FSLA checkpoint conversion or training, the `turbodiffusion-serve` TUI, or TurboT2AV audio-video generation.\n\nTurboDiffusion is a CUDA-first package. Most real generation workflows need a compatible NVIDIA GPU, model checkpoints, Wan VAE\u002Ftext encoder files, and source-layout imports. Do not start downloads, long generation, or training unless the user explicitly supplied the needed assets and budget.\n\n## First checks\n\n1. Read [repo provenance](references\u002Frepo-provenance.md) when deciding whether this skill is current for a checkout.\n2. Read [installation and backend notes](references\u002Finstallation-and-backends.md) before installing, compiling, or diagnosing CUDA\u002FSageSLA\u002Fcustom-op problems.\n3. Run [scripts\u002Fcheck_turbodiffusion_env.py](scripts\u002Fcheck_turbodiffusion_env.py) when a user asks whether their current environment can import TurboDiffusion, see CUDA, load custom ops, or use optional SageSLA.\n4. Read [model and asset catalog](references\u002Fmodel-and-asset-catalog.md) before building commands that mention VAE\u002Ftext encoder\u002FDiT checkpoints, high\u002Flow I2V checkpoints, prompt files, or TurboT2AV assets.\n5. Use [cross-cutting troubleshooting](references\u002Ftroubleshooting.md) for install\u002Fimport, source-layout, missing checkpoint, CUDA, and optional dependency failures.\n\n## Sub-skill routing\n\n| User request | Read |\n| --- | --- |\n| Build a one-shot text-to-video or image-to-video command; choose Wan2.1\u002FWan2.2 model flags; validate prompt\u002Fimage\u002Fcheckpoint arguments; avoid running models while planning. | [video-inference](sub-skills\u002Fvideo-inference\u002FSKILL.md) |\n| Launch or debug the interactive terminal server; decide T2V vs I2V TUI flags; understand slash commands and runtime-adjustable parameters. | [interactive-serving](sub-skills\u002Finteractive-serving\u002FSKILL.md) |\n| Prepare SLA\u002FrCM training commands, checkpoint conversion, checkpoint merging, safetensors-to-pth conversion, quantized checkpoint export, or data\u002Fcheckpoint layout checks. | [training-and-checkpoints](sub-skills\u002Ftraining-and-checkpoints\u002FSKILL.md) |\n| Diagnose CUDA extension build\u002Fimport, `turbo_diffusion_ops`, INT8\u002FFastNorm, `attention_type=sla` vs `sagesla`, SpargeAttn, `flash-attn`, or source-layout import quirks. | [acceleration-backends](sub-skills\u002Facceleration-backends\u002FSKILL.md) |\n| Plan TurboT2AV text-to-audio-video generation, LTX-2\u002FPixi setup, HF\u002FGemma checkpoint requirements, or TileLang\u002FSageSLA acceleration choices. | [turbot2av-extension](sub-skills\u002Fturbot2av-extension\u002FSKILL.md) |\n\n## Operating rules\n\n- Prefer the bundled command builders in sub-skill `scripts\u002F` for command construction. They render commands and validate obvious mistakes without downloading weights, starting training, or running generation.\n- For source checkouts, TurboDiffusion scripts and even the installed TUI entry point may require adding the package source directory to `PYTHONPATH` so top-level imports such as `imaginaire`, `rcm`, `serve`, and `modify_model` resolve. Capture this as a public source-layout requirement, not as a local path.\n- Treat `attention_type=sagesla` as optional-backend dependent: install SpargeAttn first or fall back to `sla`\u002F`original` when SageSLA is not available.\n- Match `--quant_linear` to quantized checkpoints. Do not add it to unquantized checkpoints on large-memory GPUs unless the user explicitly wants the quantized path.\n- Do not run full T2V\u002FI2V\u002FTurboT2AV generation unless the user provides the checkpoints, VAE\u002Ftext encoder assets, output location, GPU budget, and permission to spend time\u002FGPU memory.\n- Do not run full rCM\u002FSLA training unless the user provides checkpoint\u002Fdata roots, distributed\u002FGPU budget, logging policy, and any credentials such as WandB or Hugging Face tokens.\n\n## Minimal environment smoke\n\nUse this public smoke from an installed\u002Fsource-layout environment before deeper work:\n\n```bash\npython - \u003C\u003C'PY'\nimport torch\nprint('torch', torch.__version__, torch.version.cuda, torch.cuda.is_available())\ntry:\n    import turbo_diffusion_ops\n    print('custom ops: ok')\nexcept Exception as exc:\n    print('custom ops:', type(exc).__name__, exc)\nPY\n```\n\nFor a fuller check, run the bundled [environment diagnostic](scripts\u002Fcheck_turbodiffusion_env.py). If it reports missing `imaginaire` or `rcm`, use the source-layout `PYTHONPATH` guidance in [installation and backend notes](references\u002Finstallation-and-backends.md).\n","# TurboDiffusion\n\nUse this repo skill when a task involves TurboDiffusion, TurboWan, Wan2.1\u002FWan2.2 accelerated video generation, SLA\u002FSageSLA attention, rCM\u002FSLA checkpoint conversion or training, the `turbodiffusion-serve` TUI, or TurboT2AV audio-video generation.\n\nTurboDiffusion is a CUDA-first package. Most real generation workflows need a compatible NVIDIA GPU, model checkpoints, Wan VAE\u002Ftext encoder files, and source-layout imports. Do not start downloads, long generation, or training unless the user explicitly supplied the needed assets and budget.\n\n## First checks\n\n1. Read [repo provenance](references\u002Frepo-provenance.md) when deciding whether this skill is current for a checkout.\n2. Read [installation and backend notes](references\u002Finstallation-and-backends.md) before installing, compiling, or diagnosing CUDA\u002FSageSLA\u002Fcustom-op problems.\n3. Run [scripts\u002Fcheck_turbodiffusion_env.py](scripts\u002Fcheck_turbodiffusion_env.py) when a user asks whether their current environment can import TurboDiffusion, see CUDA, load custom ops, or use optional SageSLA.\n4. Read [model and asset catalog](references\u002Fmodel-and-asset-catalog.md) before building commands that mention VAE\u002Ftext encoder\u002FDiT checkpoints, high\u002Flow I2V checkpoints, prompt files, or TurboT2AV assets.\n5. Use [cross-cutting troubleshooting](references\u002Ftroubleshooting.md) for install\u002Fimport, source-layout, missing checkpoint, CUDA, and optional dependency failures.\n\n## Sub-skill routing\n\n| User request | Read |\n| --- | --- |\n| Build a one-shot text-to-video or image-to-video command; choose Wan2.1\u002FWan2.2 model flags; validate prompt\u002Fimage\u002Fcheckpoint arguments; avoid running models while planning. | [video-inference](sub-skills\u002Fvideo-inference\u002FSKILL.md) |\n| Launch or debug the interactive terminal server; decide T2V vs I2V TUI flags; understand slash commands and runtime-adjustable parameters. | [interactive-serving](sub-skills\u002Finteractive-serving\u002FSKILL.md) |\n| Prepare SLA\u002FrCM training commands, checkpoint conversion, checkpoint merging, safetensors-to-pth conversion, quantized checkpoint export, or data\u002Fcheckpoint layout checks. | [training-and-checkpoints](sub-skills\u002Ftraining-and-checkpoints\u002FSKILL.md) |\n| Diagnose CUDA extension build\u002Fimport, `turbo_diffusion_ops`, INT8\u002FFastNorm, `attention_type=sla` vs `sagesla`, SpargeAttn, `flash-attn`, or source-layout import quirks. | [acceleration-backends](sub-skills\u002Facceleration-backends\u002FSKILL.md) |\n| Plan TurboT2AV text-to-audio-video generation, LTX-2\u002FPixi setup, HF\u002FGemma checkpoint requirements, or TileLang\u002FSageSLA acceleration choices. | [turbot2av-extension](sub-skills\u002Fturbot2av-extension\u002FSKILL.md) |\n\n## Operating rules\n\n- Prefer the bundled command builders in sub-skill `scripts\u002F` for command construction. They render commands and validate obvious mistakes without downloading weights, starting training, or running generation.\n- For source checkouts, TurboDiffusion scripts and even the installed TUI entry point may require adding the package source directory to `PYTHONPATH` so top-level imports such as `imaginaire`, `rcm`, `serve`, and `modify_model` resolve. Capture this as a public source-layout requirement, not as a local path.\n- Treat `attention_type=sagesla` as optional-backend dependent: install SpargeAttn first or fall back to `sla`\u002F`original` when SageSLA is not available.\n- Match `--quant_linear` to quantized checkpoints. Do not add it to unquantized checkpoints on large-memory GPUs unless the user explicitly wants the quantized path.\n- Do not run full T2V\u002FI2V\u002FTurboT2AV generation unless the user provides the checkpoints, VAE\u002Ftext encoder assets, output location, GPU budget, and permission to spend time\u002FGPU memory.\n- Do not run full rCM\u002FSLA training unless the user provides checkpoint\u002Fdata roots, distributed\u002FGPU budget, logging policy, and any credentials such as WandB or Hugging Face tokens.\n\n## Minimal environment smoke\n\nUse this public smoke from an installed\u002Fsource-layout environment before deeper work:\n\n```bash\npython - \u003C\u003C'PY'\nimport torch\nprint('torch', torch.__version__, torch.version.cuda, torch.cuda.is_available())\ntry:\n    import turbo_diffusion_ops\n    print('custom ops: ok')\nexcept Exception as exc:\n    print('custom ops:', type(exc).__name__, exc)\nPY\n```\n\nFor a fuller check, run the bundled [environment diagnostic](scripts\u002Fcheck_turbodiffusion_env.py). If it reports missing `imaginaire` or `rcm`, use the source-layout `PYTHONPATH` guidance in [installation and backend notes](references\u002Finstallation-and-backends.md).\n",{"name":5,"license":23,"metadata":24,"description":7,"disable-model-invocation":26},"Apache 2.0",{"disco-role":25},"operating",true,"main","ac3fe1afa80fb9a09775ecfb2b6cc3ba850a2db6","skills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion",27,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab","2026-09-03 10:25:21",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":26},[36,39,42],{"tool":37,"command":38},"npx","npx skills add VectorSpaceLab\u002FAREX-Skill --skill turbo-diffusion",{"tool":40,"command":41},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill turbo-diffusion",{"tool":43,"command":44},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill turbo-diffusion","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002FSKILL.md",[47,53,57,61,65,69],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Finstallation-and-backends.md","REFERENCE","参考文档",4268,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002Freferences\u002Finstallation-and-backends.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Fmodel-and-asset-catalog.md",4665,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002Freferences\u002Fmodel-and-asset-catalog.md",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Frepo-provenance.md",3683,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002Freferences\u002Frepo-provenance.md",{"path":62,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":63,"sourceHtmlUrl":64},"references\u002Frepo-routing-metadata.json",516,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002Freferences\u002Frepo-routing-metadata.json",{"path":66,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":67,"sourceHtmlUrl":68},"references\u002Ftroubleshooting.md",4267,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002Freferences\u002Ftroubleshooting.md",{"path":70,"resourceType":71,"resourceTypeDesc":72,"sizeBytes":73,"sourceHtmlUrl":74},"scripts\u002Fcheck_turbodiffusion_env.py","SCRIPT","脚本",5502,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fturbo-diffusion\u002Fscripts\u002Fcheck_turbodiffusion_env.py",[76,83,90,97,104,111],{"skillNo":77,"name":78,"nameZh":6,"description":79,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":18},"01M799NYSH1R3RVAHSM2D88Y84","passnet-triton-opt","Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block\u002Fgrid\u002Fwarp 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.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-triton-opt\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":18},"01M799NS5APEC0N1FZ7XZVW6BN","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.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-solve\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":18},"01M799NNY9BKSM21V5R7QVTBJX","passnet-skill","PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under .\u002Fpass_dir\u002F, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-skill\u002FSKILL.md",[],[],{"skillNo":98,"name":99,"nameZh":6,"description":100,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":101,"categories":102,"keywords":103,"publishedAt":18},"01M799NHH6NP7Z1RK77KGXAN6S","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()\u002Freplacement_args()\u002Freplacement_func() files, when a pass \"failed to match\", or when deciding how to fuse multiple kernels\u002Fops.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-pattern-fusion\u002FSKILL.md",[],[],{"skillNo":105,"name":106,"nameZh":6,"description":107,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":108,"categories":109,"keywords":110,"publishedAt":18},"01M799ND2J4WNC2PGAZ0A06NDF","passnet-orchestrate","Round\u002Fbatch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep\u002Frevert 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.","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-orchestrate\u002FSKILL.md",[],[],{"skillNo":112,"name":113,"nameZh":6,"description":114,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":115,"categories":116,"keywords":117,"publishedAt":18},"01M799N8SBTMY02ZDG2AZF5ZWE","passnet-feedback","Fast, deterministic feedback for PassNet work: pre-flight pattern\u002Fmatch 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).","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-feedback\u002FSKILL.md",[],[]]