[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78STBBPD95B6A2MYC2H6NBV":3,"skill-same:01M78STBBPD95B6A2MYC2H6NBV":66},{"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":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":32,"install":33,"resources":45},"01M78STBBPD95B6A2MYC2H6NBV","custom-models",null,"Build or debug Asteroid custom models, blocks, registries, DSP helpers, and tracing behavior.","VectorSpaceLab\u002FAREX-Skill","VectorSpaceLab","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F180637102?v=4","01M7538C5WPSJGWF7KHR4EFZ9F","01M7538C6R2GFP53NPXTCZNBRT",351,"MIT","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: custom-models\ndescription: \"Build or debug Asteroid custom models, blocks, registries, DSP\n  helpers, and tracing behavior.\"\nmetadata:\n  disco-role: operating\ndisable-model-invocation: true\nlicense: MIT\n---\n\n# Custom models and core APIs\n\nUse this sub-skill when the user is building a new Asteroid architecture, extending a block, or debugging a low-level API rather than running a recipe or applying a pretrained model.\n\n## Typical triggers\n\n- custom `BaseModel` or `BaseEncoderMaskerDecoder` subclasses\n- `asteroid.models.get(...)` or `register_model(...)`\n- `asteroid.masknn`, `asteroid.dsp`, `asteroid.complex_nn`, or `asteroid.utils`\n- filterbanks, encoders, decoders, mask networks, or normalization layers\n- TorchScript \u002F tracing \u002F shape issues\n- `register_norm(...)`, `register_optimizer(...)`, or `prepare_parser_from_dict(...)`\n\n## What to do first\n\n1. Identify the exact block or API family the user wants to extend.\n2. Decide whether they need:\n   - a built-in model family\n   - a reusable block\n   - a custom serialization contract\n   - a tracing or shape-debugging pass\n3. Check the expected input\u002Foutput shape and the sample-rate contract before editing anything.\n\n## Standard workflow\n\n- Read `references\u002Fapi-reference.md` for the model\u002Fblock\u002Fregistry surface.\n- Read `references\u002Fjit-and-tracing.md` when TorchScript or `torch.jit.trace` is involved.\n- Read `references\u002Ftroubleshooting.md` when the issue looks like a shape, serialization, or registry mismatch.\n- Use `scripts\u002Fsmoke_building_blocks.py` to exercise representative blocks and utilities locally.\n\n## Common task families\n\n- new source-separation architectures\n- custom encoders\u002Fdecoders or filterbanks\n- mask network experiments\n- complex-number-aware layers\n- batch or source-dimension shape debugging\n- registry-based extension points for activations, norms, optimizers, or models\n\n## Built-in model families worth recognizing\n\nThe repo exposes ready-to-use model families such as:\n\n- ConvTasNet\n- DPRNNTasNet\n- DPTNet\n- LSTMTasNet\n- DeMask\n- DCUNet\n- DCCRNet\n- SuDORMRFNet \u002F SuDORMRFImprovedNet\n- FasNetTAC\n- XUMX\n\nThose are often the right starting point when the user wants to modify an existing Asteroid model rather than write a fresh stack from scratch.\n\n## Troubleshooting reminders\n\n- `get_model_args()` should be able to reconstruct the model cleanly.\n- `BaseModel.from_pretrained(...)` round-trips are a good shape\u002Fserialization smoke.\n- `BaseEncoderMaskerDecoder` expects waveform-like tensors with time last.\n- Complex-mask helpers and JIT helpers have their own shape rules.\n- If a registry lookup fails, check whether the class was actually registered and whether the name collides with an existing symbol.\n\n## Inputs to inspect\n\n- desired input shape and output shape\n- time axis, channel axis, and batch axis conventions\n- whether the user needs a built-in family, a reusable block, or a registry extension point\n- whether tracing or serialization must work in addition to eager execution\n\n## Smoke sequence\n\n1. Instantiate a tiny built-in model or block.\n2. Run a forward pass on a tiny tensor.\n3. Round-trip through `serialize()` \u002F `from_pretrained()` if a model is involved.\n4. Trace the same block if the request mentions JIT or export.\n5. Confirm any registry or helper function the user cares about.\n\n## What to avoid\n\n- Do not hand-wave shape conventions.\n- Do not treat a successful eager forward pass as proof of tracing correctness.\n- Do not rely on the original repo checkout in runtime instructions.\n- Do not blur custom-model work with recipe training unless the task truly spans both.\n\n## Common model-building signals\n\n- `filterbank`, `encoder`, `decoder`, or `masker` usually point to architecture assembly.\n- `complex` and `beamforming` usually point to DSP or complex-valued helpers.\n- `register_*` almost always means a registry extension point.\n- `torch.jit` or `trace` means the tracing-safe path should be considered explicitly.\n\n## Good questions to ask when unclear\n\n- Is this a new model or a modification of an existing family?\n- Which dimensions should be preserved in the output?\n- Does the user need a round-trip serialization contract?\n- Should the answer focus on eager execution, tracing, or both?\n","# Custom models and core APIs\n\nUse this sub-skill when the user is building a new Asteroid architecture, extending a block, or debugging a low-level API rather than running a recipe or applying a pretrained model.\n\n## Typical triggers\n\n- custom `BaseModel` or `BaseEncoderMaskerDecoder` subclasses\n- `asteroid.models.get(...)` or `register_model(...)`\n- `asteroid.masknn`, `asteroid.dsp`, `asteroid.complex_nn`, or `asteroid.utils`\n- filterbanks, encoders, decoders, mask networks, or normalization layers\n- TorchScript \u002F tracing \u002F shape issues\n- `register_norm(...)`, `register_optimizer(...)`, or `prepare_parser_from_dict(...)`\n\n## What to do first\n\n1. Identify the exact block or API family the user wants to extend.\n2. Decide whether they need:\n   - a built-in model family\n   - a reusable block\n   - a custom serialization contract\n   - a tracing or shape-debugging pass\n3. Check the expected input\u002Foutput shape and the sample-rate contract before editing anything.\n\n## Standard workflow\n\n- Read `references\u002Fapi-reference.md` for the model\u002Fblock\u002Fregistry surface.\n- Read `references\u002Fjit-and-tracing.md` when TorchScript or `torch.jit.trace` is involved.\n- Read `references\u002Ftroubleshooting.md` when the issue looks like a shape, serialization, or registry mismatch.\n- Use `scripts\u002Fsmoke_building_blocks.py` to exercise representative blocks and utilities locally.\n\n## Common task families\n\n- new source-separation architectures\n- custom encoders\u002Fdecoders or filterbanks\n- mask network experiments\n- complex-number-aware layers\n- batch or source-dimension shape debugging\n- registry-based extension points for activations, norms, optimizers, or models\n\n## Built-in model families worth recognizing\n\nThe repo exposes ready-to-use model families such as:\n\n- ConvTasNet\n- DPRNNTasNet\n- DPTNet\n- LSTMTasNet\n- DeMask\n- DCUNet\n- DCCRNet\n- SuDORMRFNet \u002F SuDORMRFImprovedNet\n- FasNetTAC\n- XUMX\n\nThose are often the right starting point when the user wants to modify an existing Asteroid model rather than write a fresh stack from scratch.\n\n## Troubleshooting reminders\n\n- `get_model_args()` should be able to reconstruct the model cleanly.\n- `BaseModel.from_pretrained(...)` round-trips are a good shape\u002Fserialization smoke.\n- `BaseEncoderMaskerDecoder` expects waveform-like tensors with time last.\n- Complex-mask helpers and JIT helpers have their own shape rules.\n- If a registry lookup fails, check whether the class was actually registered and whether the name collides with an existing symbol.\n\n## Inputs to inspect\n\n- desired input shape and output shape\n- time axis, channel axis, and batch axis conventions\n- whether the user needs a built-in family, a reusable block, or a registry extension point\n- whether tracing or serialization must work in addition to eager execution\n\n## Smoke sequence\n\n1. Instantiate a tiny built-in model or block.\n2. Run a forward pass on a tiny tensor.\n3. Round-trip through `serialize()` \u002F `from_pretrained()` if a model is involved.\n4. Trace the same block if the request mentions JIT or export.\n5. Confirm any registry or helper function the user cares about.\n\n## What to avoid\n\n- Do not hand-wave shape conventions.\n- Do not treat a successful eager forward pass as proof of tracing correctness.\n- Do not rely on the original repo checkout in runtime instructions.\n- Do not blur custom-model work with recipe training unless the task truly spans both.\n\n## Common model-building signals\n\n- `filterbank`, `encoder`, `decoder`, or `masker` usually point to architecture assembly.\n- `complex` and `beamforming` usually point to DSP or complex-valued helpers.\n- `register_*` almost always means a registry extension point.\n- `torch.jit` or `trace` means the tracing-safe path should be considered explicitly.\n\n## Good questions to ask when unclear\n\n- Is this a new model or a modification of an existing family?\n- Which dimensions should be preserved in the output?\n- Does the user need a round-trip serialization contract?\n- Should the answer focus on eager execution, tracing, or both?\n",{"name":5,"license":14,"metadata":23,"description":7,"disable-model-invocation":25},{"disco-role":24},"operating",true,"main","ac3fe1afa80fb9a09775ecfb2b6cc3ba850a2db6","skills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models",27,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab","2026-09-03 10:25:21",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":25},[35,38,41],{"tool":36,"command":37},"npx","npx skills add VectorSpaceLab\u002FAREX-Skill --skill custom-models",{"tool":39,"command":40},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill custom-models",{"tool":42,"command":43},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill custom-models","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models\u002FSKILL.md",[46,52,56,60],{"path":47,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":51},"references\u002Fapi-reference.md","REFERENCE","参考文档",2715,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models\u002Freferences\u002Fapi-reference.md",{"path":53,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":54,"sourceHtmlUrl":55},"references\u002Fjit-and-tracing.md",1158,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models\u002Freferences\u002Fjit-and-tracing.md",{"path":57,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Ftroubleshooting.md",1857,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models\u002Freferences\u002Ftroubleshooting.md",{"path":61,"resourceType":62,"resourceTypeDesc":63,"sizeBytes":64,"sourceHtmlUrl":65},"scripts\u002Fsmoke_building_blocks.py","SCRIPT","脚本",2010,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fasteroid\u002Fsub-skills\u002Fcustom-models\u002Fscripts\u002Fsmoke_building_blocks.py",[67,75,82,89,96,103],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":71,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"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.","Apache-2.0","https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Ftask-oriented\u002FPassNet\u002Fskills\u002Fpassnet-triton-opt\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":71,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"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":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":71,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"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":90,"name":91,"nameZh":6,"description":92,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":71,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"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":97,"name":98,"nameZh":6,"description":99,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":71,"licenseFileUrl":6,"sourceHtmlUrl":100,"categories":101,"keywords":102,"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":104,"name":105,"nameZh":6,"description":106,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":71,"licenseFileUrl":6,"sourceHtmlUrl":107,"categories":108,"keywords":109,"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",[],[]]