[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M796EQC54C2WMY8S48SMDM2X":3,"skill-same:01M796EQC54C2WMY8S48SMDM2X":74},{"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},"01M796EQC54C2WMY8S48SMDM2X","collectors-and-replay",null,"Collect rollouts, select collector topologies, configure replay buffers, and debug TorchRL data movement.","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\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: collectors-and-replay\ndescription: \"Collect rollouts, select collector topologies, configure replay\n  buffers, and debug TorchRL data movement.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: MIT\n---\n\n# TorchRL collectors and replay\n\nUse this sub-skill when a task involves TorchRL rollout collection, evaluation rollouts, replay-buffer composition, sequence or slice sampling, prioritized replay, generation-safe record updates, memmap storage, checkpointing, HER, or optional service-backed replay.\n\n## Route here for\n\n- `Collector`, `AsyncCollector`, `MultiCollector`, `Evaluator`, `frames_per_batch`, `total_frames`, `trajs_per_batch`, `sync`, `backend`, `backend_options`, `num_collectors`, `update_policy_weights_`, and weight-update schemes.\n- Direct, local process, Ray, RPC, distributed, and Submitit collector topology selection.\n- `ReplayBuffer`, `TensorDictReplayBuffer`, `TensorDictPrioritizedReplayBuffer`, `LazyTensorStorage`, `LazyMemmapStorage`, `SliceSampler`, `Sequence`, `PrioritizedSampler`, `TensorDictRoundRobinWriter`, `update_tensordict_priority`, and `update_if_present`.\n- Collector-to-replay integration, explicit `device` \u002F `storing_device` \u002F `policy_device` \u002F `env_device` decisions, and debugging TensorDict movement across collection and storage.\n\n## Route out instead\n\n- Environment construction, specs, transforms, Gym\u002FGymnasium wrappers, and `step_mdp` layout questions: use `envs-and-transforms`.\n- Actor, critic, recurrent policy, distribution, and TensorDictModule construction: use `modules-and-policies`.\n- Loss modules, trainers, target updates, and algorithm optimization loops: use `objectives-and-training`.\n- Generic TorchRL service registry, LLM collectors, VLA schemas, rendering, or serving backends: use `llm-vla-and-services`.\n\n## Operating references\n\n1. Read [collector-workflows.md](references\u002Fcollector-workflows.md) for collector topology, sync\u002Fasync choices, explicit devices, Evaluator, and collector-to-replay patterns.\n2. Read [replay-buffer-workflows.md](references\u002Freplay-buffer-workflows.md) for buffer composition, storages, samplers, writers, transforms, sequence units, priorities, HER, memmap, checkpointing, and service-backed replay.\n3. Read [api-reference.md](references\u002Fapi-reference.md) for verified signatures and import paths.\n4. Read [troubleshooting.md](references\u002Ftroubleshooting.md) when collection, storage, sampling, priority updates, process workers, memmap cleanup, or optional Ray services fail.\n\n## Bundled smoke helpers\n\n- [scripts\u002Fsmoke_collector.py](scripts\u002Fsmoke_collector.py): CPU-only direct collector, explicit devices, direct collector-to-replay integration, and synchronous Evaluator smoke.\n- [scripts\u002Fsmoke_replay_buffer.py](scripts\u002Fsmoke_replay_buffer.py): CPU-only TensorDict replay, LazyTensor\u002FLazyMemmap storage, prioritized replay, `Sequence` sample unit masks, checkpoint round trip, and generation-safe updates.\n\nRun these helpers from any working directory after TorchRL and TensorDict are importable. They do not download assets, launch distributed services, or train models.\n\n## Evidence basis\n\nThis runtime guidance was distilled from relative TorchRL evidence covering `torchrl\u002Fcollectors\u002F`, `torchrl\u002Fdata\u002Freplay_buffers\u002F`, collector and replay reference pages, collector\u002Fevaluator\u002Freplay tutorials, selected collector and replay examples, selected `test\u002Fcollectors\u002F` and `test\u002Frb\u002F` behavior, and installed API probes. Optional Ray, RPC, Submitit, CUDA, simulator, and service-backed behaviors are documented as optional or reference-only unless a future task verifies those backends explicitly.\n","# TorchRL collectors and replay\n\nUse this sub-skill when a task involves TorchRL rollout collection, evaluation rollouts, replay-buffer composition, sequence or slice sampling, prioritized replay, generation-safe record updates, memmap storage, checkpointing, HER, or optional service-backed replay.\n\n## Route here for\n\n- `Collector`, `AsyncCollector`, `MultiCollector`, `Evaluator`, `frames_per_batch`, `total_frames`, `trajs_per_batch`, `sync`, `backend`, `backend_options`, `num_collectors`, `update_policy_weights_`, and weight-update schemes.\n- Direct, local process, Ray, RPC, distributed, and Submitit collector topology selection.\n- `ReplayBuffer`, `TensorDictReplayBuffer`, `TensorDictPrioritizedReplayBuffer`, `LazyTensorStorage`, `LazyMemmapStorage`, `SliceSampler`, `Sequence`, `PrioritizedSampler`, `TensorDictRoundRobinWriter`, `update_tensordict_priority`, and `update_if_present`.\n- Collector-to-replay integration, explicit `device` \u002F `storing_device` \u002F `policy_device` \u002F `env_device` decisions, and debugging TensorDict movement across collection and storage.\n\n## Route out instead\n\n- Environment construction, specs, transforms, Gym\u002FGymnasium wrappers, and `step_mdp` layout questions: use `envs-and-transforms`.\n- Actor, critic, recurrent policy, distribution, and TensorDictModule construction: use `modules-and-policies`.\n- Loss modules, trainers, target updates, and algorithm optimization loops: use `objectives-and-training`.\n- Generic TorchRL service registry, LLM collectors, VLA schemas, rendering, or serving backends: use `llm-vla-and-services`.\n\n## Operating references\n\n1. Read [collector-workflows.md](references\u002Fcollector-workflows.md) for collector topology, sync\u002Fasync choices, explicit devices, Evaluator, and collector-to-replay patterns.\n2. Read [replay-buffer-workflows.md](references\u002Freplay-buffer-workflows.md) for buffer composition, storages, samplers, writers, transforms, sequence units, priorities, HER, memmap, checkpointing, and service-backed replay.\n3. Read [api-reference.md](references\u002Fapi-reference.md) for verified signatures and import paths.\n4. Read [troubleshooting.md](references\u002Ftroubleshooting.md) when collection, storage, sampling, priority updates, process workers, memmap cleanup, or optional Ray services fail.\n\n## Bundled smoke helpers\n\n- [scripts\u002Fsmoke_collector.py](scripts\u002Fsmoke_collector.py): CPU-only direct collector, explicit devices, direct collector-to-replay integration, and synchronous Evaluator smoke.\n- [scripts\u002Fsmoke_replay_buffer.py](scripts\u002Fsmoke_replay_buffer.py): CPU-only TensorDict replay, LazyTensor\u002FLazyMemmap storage, prioritized replay, `Sequence` sample unit masks, checkpoint round trip, and generation-safe updates.\n\nRun these helpers from any working directory after TorchRL and TensorDict are importable. They do not download assets, launch distributed services, or train models.\n\n## Evidence basis\n\nThis runtime guidance was distilled from relative TorchRL evidence covering `torchrl\u002Fcollectors\u002F`, `torchrl\u002Fdata\u002Freplay_buffers\u002F`, collector and replay reference pages, collector\u002Fevaluator\u002Freplay tutorials, selected collector and replay examples, selected `test\u002Fcollectors\u002F` and `test\u002Frb\u002F` behavior, and installed API probes. Optional Ray, RPC, Submitit, CUDA, simulator, and service-backed behaviors are documented as optional or reference-only unless a future task verifies those backends explicitly.\n",{"name":5,"license":14,"metadata":23,"description":7,"disable-model-invocation":25},{"disco-role":24},"operating",true,"main","ac3fe1afa80fb9a09775ecfb2b6cc3ba850a2db6","skills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay",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 collectors-and-replay",{"tool":39,"command":40},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill collectors-and-replay",{"tool":42,"command":43},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill collectors-and-replay","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002FSKILL.md",[46,52,56,60,64,70],{"path":47,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":51},"references\u002Fapi-reference.md","REFERENCE","参考文档",9280,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002Freferences\u002Fapi-reference.md",{"path":53,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":54,"sourceHtmlUrl":55},"references\u002Fcollector-workflows.md",9608,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002Freferences\u002Fcollector-workflows.md",{"path":57,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":58,"sourceHtmlUrl":59},"references\u002Freplay-buffer-workflows.md",11537,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002Freferences\u002Freplay-buffer-workflows.md",{"path":61,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":62,"sourceHtmlUrl":63},"references\u002Ftroubleshooting.md",9234,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002Freferences\u002Ftroubleshooting.md",{"path":65,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":68,"sourceHtmlUrl":69},"scripts\u002Fsmoke_collector.py","SCRIPT","脚本",3493,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002Fscripts\u002Fsmoke_collector.py",{"path":71,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":72,"sourceHtmlUrl":73},"scripts\u002Fsmoke_replay_buffer.py",5087,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Ftorchrl\u002Fsub-skills\u002Fcollectors-and-replay\u002Fscripts\u002Fsmoke_replay_buffer.py",[75,83,90,97,104,111],{"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":79,"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.","Apache-2.0","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":79,"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":79,"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":79,"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":79,"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":79,"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",[],[]]