[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78YPFR0RTYGDKJH2ZN3A82M":3,"skill-same:01M78YPFR0RTYGDKJH2ZN3A82M":67},{"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},"01M78YPFR0RTYGDKJH2ZN3A82M","distributed-tuning",null,"Routes KerasTuner's chief\u002Fworker Oracle coordination over gRPC using the KERASTUNER_ORACLE_IP, KERASTUNER_ORACLE_PORT, and KERASTUNER_TUNER_ID environment contract.","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\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: distributed-tuning\nid: distributed-tuning\ndescription: \"Routes KerasTuner's chief\u002Fworker Oracle coordination over gRPC\n  using the KERASTUNER_ORACLE_IP, KERASTUNER_ORACLE_PORT, and\n  KERASTUNER_TUNER_ID environment contract.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# Distributed Tuning\n\nUse this sub-skill when a KerasTuner search must coordinate multiple tuner\nprocesses through a chief Oracle, or when a `KERASTUNER_ORACLE_*`\u002F\n`KERASTUNER_TUNER_ID` deployment is failing.\n\n## Scope\n\n- Configure the three distributed Oracle environment variables as one unit.\n- Assign chief and worker roles, launch the gRPC service, and diagnose network\n  and shared-storage prerequisites.\n- Understand OracleClient request, status, reporting, and shutdown behavior.\n- Keep distributed Oracle coordination separate from the model\u002Fsearch-space\n  design handled by the main KerasTuner workflow.\n\nDo not use this route for a local single-process search, CloudOracle-managed\ndistribution, or general `HyperParameters`\u002Ftuner selection. In particular,\nthe `Tuner` public contract supports only **single-worker** distribution\nstrategies; Oracle-level parallel tuning over its chief\u002Fworker RPC protocol is\nnot a promise that a multi-worker `distribution_strategy` is supported.\n\n## Route in this order\n\n1. Read [api-reference.md](references\u002Fapi-reference.md) for the verified\n   environment contract, role detection, RPC surface, and reporting semantics.\n2. From the skill root, run `python sub-skills\u002Fdistributed-tuning\u002Fscripts\u002Fcheck_distributed_env.py` before importing or constructing a\n   distributed tuner. It only inspects environment variables; it never starts\n   a server or opens a socket.\n3. Follow [workflows.md](references\u002Fworkflows.md) for the launch sequence,\n   shared project directory, status handling, and orderly shutdown.\n4. Use [troubleshooting.md](references\u002Ftroubleshooting.md) for missing\n   variables, accidental roles, connection stalls, filesystem divergence, and\n   cleanup failures.\n\n## Non-negotiable launch contract\n\nSet all of these together on every participating process:\n\n- `KERASTUNER_ORACLE_IP`: the address used both to bind the chief service and\n  to reach it from workers.\n- `KERASTUNER_ORACLE_PORT`: the same available TCP port for the chief service.\n- `KERASTUNER_TUNER_ID`: a deliberate identifier, unique at the independent\n  tuner\u002Fgroup coordination scope.\n\nAn ID containing the case-sensitive substring `chief` is treated as the chief\nside. The chief starts `OracleServicer`; other IDs construct `OracleClient` and\nproxy Oracle calls. A partial environment is invalid: KerasTuner raises\n`RuntimeError` when the IP is set without the port or tuner ID. The bundled\nchecker is intentionally stricter and treats empty\u002Fwhitespace values as\nmissing, including port\u002FID values supplied without an IP; do not bypass this\npreflight because runtime role detection is key-presence based.\n\nUse a reachable non-loopback address for different hosts, an open and reserved\nport, the same `directory`\u002F`project_name` on a shared writable filesystem, and\nan environment where the `grpcio` runtime dependency is installed. The RPC\ntransport is insecure gRPC, so keep the address on a trusted network or behind\nan appropriate network control; this protocol does not provide TLS.\n\n## Role and lifecycle guardrails\n\n- Start the same tuner program with the chief environment on exactly one\n  process\u002Fgroup. Its `search()` path saves the Oracle and starts the\n  `OracleServicer` server, which otherwise would not block the process.\n- Start workers with the same Oracle IP\u002Fport and non-chief IDs. Clients wait\n  for the chief service with `wait_for_ready=True` for up to 60 minutes per\n  RPC; this tolerates startup races but does not repair a wrong address,\n  blocked port, or an absent chief.\n- Give independent tuner processes\u002Fgroups distinct IDs. If a training backend\n  groups processes into one multi-worker strategy, preserve that backend's\n  required ID convention rather than inventing a second Oracle group.\n- Keep the project directory shared and writable until all workers have\n  finished. Do not let multiple processes independently delete it with\n  `overwrite=True`.\n- Let workers observe `STOPPED` and exit their search loop; do not kill the\n  chief before ongoing trials and registered tuner IDs have drained. The\n  chief's server wait loop exits after its stop signal and drain condition.\n\nThe test helper `keras_tuner.test_utils.mock_distribute` is evidence for the\nenvironment protocol only. It patches `os.environ`, creates threads, and\npicks a port for tests; it is not a production launcher or runtime dependency.\n\n## Verification focus\n\nA useful check must show: an entirely absent environment remains valid ordinary\nlocal mode, no partial environment is accepted as ordinary local mode, and a\ncomplete set is accepted without a socket or server; a chief ID is not confused\nwith a worker ID; and all participants refer to one reachable endpoint and one\nshared project directory. Do not use the mock test helper as proof that a real\ncross-host network or filesystem is configured.\n","# Distributed Tuning\n\nUse this sub-skill when a KerasTuner search must coordinate multiple tuner\nprocesses through a chief Oracle, or when a `KERASTUNER_ORACLE_*`\u002F\n`KERASTUNER_TUNER_ID` deployment is failing.\n\n## Scope\n\n- Configure the three distributed Oracle environment variables as one unit.\n- Assign chief and worker roles, launch the gRPC service, and diagnose network\n  and shared-storage prerequisites.\n- Understand OracleClient request, status, reporting, and shutdown behavior.\n- Keep distributed Oracle coordination separate from the model\u002Fsearch-space\n  design handled by the main KerasTuner workflow.\n\nDo not use this route for a local single-process search, CloudOracle-managed\ndistribution, or general `HyperParameters`\u002Ftuner selection. In particular,\nthe `Tuner` public contract supports only **single-worker** distribution\nstrategies; Oracle-level parallel tuning over its chief\u002Fworker RPC protocol is\nnot a promise that a multi-worker `distribution_strategy` is supported.\n\n## Route in this order\n\n1. Read [api-reference.md](references\u002Fapi-reference.md) for the verified\n   environment contract, role detection, RPC surface, and reporting semantics.\n2. From the skill root, run `python sub-skills\u002Fdistributed-tuning\u002Fscripts\u002Fcheck_distributed_env.py` before importing or constructing a\n   distributed tuner. It only inspects environment variables; it never starts\n   a server or opens a socket.\n3. Follow [workflows.md](references\u002Fworkflows.md) for the launch sequence,\n   shared project directory, status handling, and orderly shutdown.\n4. Use [troubleshooting.md](references\u002Ftroubleshooting.md) for missing\n   variables, accidental roles, connection stalls, filesystem divergence, and\n   cleanup failures.\n\n## Non-negotiable launch contract\n\nSet all of these together on every participating process:\n\n- `KERASTUNER_ORACLE_IP`: the address used both to bind the chief service and\n  to reach it from workers.\n- `KERASTUNER_ORACLE_PORT`: the same available TCP port for the chief service.\n- `KERASTUNER_TUNER_ID`: a deliberate identifier, unique at the independent\n  tuner\u002Fgroup coordination scope.\n\nAn ID containing the case-sensitive substring `chief` is treated as the chief\nside. The chief starts `OracleServicer`; other IDs construct `OracleClient` and\nproxy Oracle calls. A partial environment is invalid: KerasTuner raises\n`RuntimeError` when the IP is set without the port or tuner ID. The bundled\nchecker is intentionally stricter and treats empty\u002Fwhitespace values as\nmissing, including port\u002FID values supplied without an IP; do not bypass this\npreflight because runtime role detection is key-presence based.\n\nUse a reachable non-loopback address for different hosts, an open and reserved\nport, the same `directory`\u002F`project_name` on a shared writable filesystem, and\nan environment where the `grpcio` runtime dependency is installed. The RPC\ntransport is insecure gRPC, so keep the address on a trusted network or behind\nan appropriate network control; this protocol does not provide TLS.\n\n## Role and lifecycle guardrails\n\n- Start the same tuner program with the chief environment on exactly one\n  process\u002Fgroup. Its `search()` path saves the Oracle and starts the\n  `OracleServicer` server, which otherwise would not block the process.\n- Start workers with the same Oracle IP\u002Fport and non-chief IDs. Clients wait\n  for the chief service with `wait_for_ready=True` for up to 60 minutes per\n  RPC; this tolerates startup races but does not repair a wrong address,\n  blocked port, or an absent chief.\n- Give independent tuner processes\u002Fgroups distinct IDs. If a training backend\n  groups processes into one multi-worker strategy, preserve that backend's\n  required ID convention rather than inventing a second Oracle group.\n- Keep the project directory shared and writable until all workers have\n  finished. Do not let multiple processes independently delete it with\n  `overwrite=True`.\n- Let workers observe `STOPPED` and exit their search loop; do not kill the\n  chief before ongoing trials and registered tuner IDs have drained. The\n  chief's server wait loop exits after its stop signal and drain condition.\n\nThe test helper `keras_tuner.test_utils.mock_distribute` is evidence for the\nenvironment protocol only. It patches `os.environ`, creates threads, and\npicks a port for tests; it is not a production launcher or runtime dependency.\n\n## Verification focus\n\nA useful check must show: an entirely absent environment remains valid ordinary\nlocal mode, no partial environment is accepted as ordinary local mode, and a\ncomplete set is accepted without a socket or server; a chief ID is not confused\nwith a worker ID; and all participants refer to one reachable endpoint and one\nshared project directory. Do not use the mock test helper as proof that a real\ncross-host network or filesystem is configured.\n",{"id":5,"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\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning",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 distributed-tuning",{"tool":40,"command":41},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill distributed-tuning",{"tool":43,"command":44},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill distributed-tuning","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning\u002FSKILL.md",[47,53,57,61],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Fapi-reference.md","REFERENCE","参考文档",6642,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning\u002Freferences\u002Fapi-reference.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Ftroubleshooting.md",6765,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning\u002Freferences\u002Ftroubleshooting.md",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Fworkflows.md",7408,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning\u002Freferences\u002Fworkflows.md",{"path":62,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":65,"sourceHtmlUrl":66},"scripts\u002Fcheck_distributed_env.py","SCRIPT","脚本",2719,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Fkeras-tuner\u002Fsub-skills\u002Fdistributed-tuning\u002Fscripts\u002Fcheck_distributed_env.py",[68,75,82,89,96,103],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"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.","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":14,"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":14,"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":14,"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":14,"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":14,"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",[],[]]