[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78WKGSVFQMQ3M2YATPW9NMP":3,"skill-same:01M78WKGSVFQMQ3M2YATPW9NMP":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},"01M78WKGSVFQMQ3M2YATPW9NMP","easy-r1",null,"Use EasyR1 for multimodal LLM RL post-training, dataset\u002Freward preparation, core API debugging, and checkpoint export workflows.","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\u002Feasy-r1\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: easy-r1\ndescription: \"Use EasyR1 for multimodal LLM RL post-training, dataset\u002Freward\n  preparation, core API debugging, and checkpoint export workflows.\"\ndisable-model-invocation: true\nmetadata:\n  disco-role: operating\nlicense: Apache 2.0\n---\n\n# EasyR1\n\nUse this repo skill when a task names **EasyR1**, the `verl` package from EasyR1, or a multimodal LLM RL\u002Fpost-training workflow that looks like EasyR1: GRPO\u002FDAPO\u002FReinforce++\u002FReMax\u002FRLOO\u002FGSPO\u002FCISPO\u002FSAPO training, Ray + FSDP + vLLM rollout, Qwen\u002FQwen-VL post-training, EasyR1 reward functions, EasyR1 checkpoint merging, or `DataProto` debugging.\n\nThis skill is self-contained operating guidance distilled from the EasyR1 repository. It does not require reopening the repository checkout. Full training still requires an EasyR1-compatible CUDA runtime with Ray, vLLM, flash-attn, model weights, datasets, and enough GPU memory.\n\n## First checks\n\n- Read [references\u002Frepo-provenance.md](references\u002Frepo-provenance.md) before deciding whether this skill is current for a checkout or whether to refresh it.\n- Read [references\u002Fsetup-and-troubleshooting.md](references\u002Fsetup-and-troubleshooting.md) for install\u002Fruntime choices, Docker\u002Fcontainer guidance, CUDA\u002FvLLM\u002Fflash-attn limits, and common cross-cutting failures.\n- Run [scripts\u002Feasyr1_env_check.py](scripts\u002Feasyr1_env_check.py) when you need a safe local environment\u002Fimport\u002Fbackend check. It does not start Ray, download models, or run training.\n\n## Route by task\n\n### Training workflows\n\nLoad [sub-skills\u002Ftraining-workflows\u002FSKILL.md](sub-skills\u002Ftraining-workflows\u002FSKILL.md) when the task is to configure, lint, explain, or launch `python -m verl.trainer.main` jobs. It covers EasyR1 config hierarchy, OmegaConf CLI overrides, algorithms and loss types, Ray\u002FFSDP\u002FvLLM runtime assumptions, LoRA, logging, validation, checkpoint save\u002Fresume, multi-node setup, and training troubleshooting.\n\nTypical triggers:\n\n- \"Run EasyR1 GRPO\u002FDAPO\u002FGSPO\u002FCISPO\u002FSAPO on my dataset.\"\n- \"Build or validate an EasyR1 config.\"\n- \"Convert an example shell launch to a safer command.\"\n- \"Explain `worker.rollout.tensor_parallel_size`, `algorithm.online_filtering`, LoRA, or GPU memory settings.\"\n\n### Data and rewards\n\nLoad [sub-skills\u002Fdata-and-rewards\u002FSKILL.md](sub-skills\u002Fdata-and-rewards\u002FSKILL.md) when the task is about dataset rows, prompt templates, image\u002Fvideo columns, reward functions, `worker.reward.reward_function`, batch\u002Fsequential reward modes, DAPO overlong penalties, R1-V\u002Fmath\u002FAndroid GUI reward patterns, or validating reward outputs.\n\nTypical triggers:\n\n- \"Prepare a custom text\u002FVL dataset for EasyR1.\"\n- \"Write an EasyR1 reward function with `overall`, `accuracy`, or `accuracy_normalized`.\"\n- \"Debug image-token mismatch, overlong prompts, or missing reward keys.\"\n\n### Core APIs\n\nLoad [sub-skills\u002Fcore-apis\u002FSKILL.md](sub-skills\u002Fcore-apis\u002FSKILL.md) when the task involves EasyR1 support APIs rather than a full job launch: `DataProto`, tensor\u002Fnon-tensor batches, padding\u002Funpadding, dynamic sequence-length batching, `compute_policy_loss`, advantage\u002FKL helpers, logger\u002Ftracker helpers, or native CPU API smoke checks.\n\nTypical triggers:\n\n- \"Debug a `DataProto.union` or chunk\u002Fsplit error.\"\n- \"Use EasyR1 core algorithm functions directly.\"\n- \"Explain dynamic batching restore order or GRPO grouped rollout assertions.\"\n\n### Checkpoint export\n\nLoad [sub-skills\u002Fcheckpoint-export\u002FSKILL.md](sub-skills\u002Fcheckpoint-export\u002FSKILL.md) when the task is to inspect EasyR1 actor checkpoint directories or convert actor shards to Hugging Face format. It covers `model_world_size_\u003CN>_rank_\u003CR>.pt` shard patterns, the `huggingface\u002F` metadata directory, generation config preservation, LoRA merge requirements, upload caveats, and safe preflight inspection.\n\nTypical triggers:\n\n- \"Merge an EasyR1 actor checkpoint to Hugging Face format.\"\n- \"Preflight a checkpoint before running the model merger.\"\n- \"Explain missing rank shards, unsupported DTensor placement, or LoRA base-model errors.\"\n\n## Minimal install\u002Fimport shape\n\nFor users setting up EasyR1 itself, prefer the project-documented CUDA container or an equivalent Python 3.9+ environment with compatible PyTorch, Ray, vLLM, flash-attn, transformers, and model\u002Fdataset access. A minimal import check is:\n\n```bash\npython - \u003C\u003C'PY'\nimport verl\nfrom verl.protocol import DataProto\nfrom verl.trainer.core_algos import compute_kl\nprint(\"verl\", getattr(verl, \"__version__\", \"unknown\"))\nprint(\"DataProto\", DataProto.__name__)\nprint(\"compute_kl\", compute_kl.__name__)\nPY\n```\n\nIf `flash_attn`, `vllm`, CUDA, or Ray imports fail, use the setup\u002Ftroubleshooting reference before treating EasyR1 training as ready. CPU-safe API checks do not prove the full training runtime.\n\n## Boundaries\n\nDo not use this skill as a generic TRL\u002FOpenRLHF\u002FLlamaFactory manual. Use it when EasyR1-specific config keys, scripts, dataset\u002Freward contracts, `verl` APIs, or checkpoint layout matter. For ordinary supervised fine-tuning or inference workflows that EasyR1 explicitly does not provide, route to a more suitable LLM fine-tuning or serving skill.\n","# EasyR1\n\nUse this repo skill when a task names **EasyR1**, the `verl` package from EasyR1, or a multimodal LLM RL\u002Fpost-training workflow that looks like EasyR1: GRPO\u002FDAPO\u002FReinforce++\u002FReMax\u002FRLOO\u002FGSPO\u002FCISPO\u002FSAPO training, Ray + FSDP + vLLM rollout, Qwen\u002FQwen-VL post-training, EasyR1 reward functions, EasyR1 checkpoint merging, or `DataProto` debugging.\n\nThis skill is self-contained operating guidance distilled from the EasyR1 repository. It does not require reopening the repository checkout. Full training still requires an EasyR1-compatible CUDA runtime with Ray, vLLM, flash-attn, model weights, datasets, and enough GPU memory.\n\n## First checks\n\n- Read [references\u002Frepo-provenance.md](references\u002Frepo-provenance.md) before deciding whether this skill is current for a checkout or whether to refresh it.\n- Read [references\u002Fsetup-and-troubleshooting.md](references\u002Fsetup-and-troubleshooting.md) for install\u002Fruntime choices, Docker\u002Fcontainer guidance, CUDA\u002FvLLM\u002Fflash-attn limits, and common cross-cutting failures.\n- Run [scripts\u002Feasyr1_env_check.py](scripts\u002Feasyr1_env_check.py) when you need a safe local environment\u002Fimport\u002Fbackend check. It does not start Ray, download models, or run training.\n\n## Route by task\n\n### Training workflows\n\nLoad [sub-skills\u002Ftraining-workflows\u002FSKILL.md](sub-skills\u002Ftraining-workflows\u002FSKILL.md) when the task is to configure, lint, explain, or launch `python -m verl.trainer.main` jobs. It covers EasyR1 config hierarchy, OmegaConf CLI overrides, algorithms and loss types, Ray\u002FFSDP\u002FvLLM runtime assumptions, LoRA, logging, validation, checkpoint save\u002Fresume, multi-node setup, and training troubleshooting.\n\nTypical triggers:\n\n- \"Run EasyR1 GRPO\u002FDAPO\u002FGSPO\u002FCISPO\u002FSAPO on my dataset.\"\n- \"Build or validate an EasyR1 config.\"\n- \"Convert an example shell launch to a safer command.\"\n- \"Explain `worker.rollout.tensor_parallel_size`, `algorithm.online_filtering`, LoRA, or GPU memory settings.\"\n\n### Data and rewards\n\nLoad [sub-skills\u002Fdata-and-rewards\u002FSKILL.md](sub-skills\u002Fdata-and-rewards\u002FSKILL.md) when the task is about dataset rows, prompt templates, image\u002Fvideo columns, reward functions, `worker.reward.reward_function`, batch\u002Fsequential reward modes, DAPO overlong penalties, R1-V\u002Fmath\u002FAndroid GUI reward patterns, or validating reward outputs.\n\nTypical triggers:\n\n- \"Prepare a custom text\u002FVL dataset for EasyR1.\"\n- \"Write an EasyR1 reward function with `overall`, `accuracy`, or `accuracy_normalized`.\"\n- \"Debug image-token mismatch, overlong prompts, or missing reward keys.\"\n\n### Core APIs\n\nLoad [sub-skills\u002Fcore-apis\u002FSKILL.md](sub-skills\u002Fcore-apis\u002FSKILL.md) when the task involves EasyR1 support APIs rather than a full job launch: `DataProto`, tensor\u002Fnon-tensor batches, padding\u002Funpadding, dynamic sequence-length batching, `compute_policy_loss`, advantage\u002FKL helpers, logger\u002Ftracker helpers, or native CPU API smoke checks.\n\nTypical triggers:\n\n- \"Debug a `DataProto.union` or chunk\u002Fsplit error.\"\n- \"Use EasyR1 core algorithm functions directly.\"\n- \"Explain dynamic batching restore order or GRPO grouped rollout assertions.\"\n\n### Checkpoint export\n\nLoad [sub-skills\u002Fcheckpoint-export\u002FSKILL.md](sub-skills\u002Fcheckpoint-export\u002FSKILL.md) when the task is to inspect EasyR1 actor checkpoint directories or convert actor shards to Hugging Face format. It covers `model_world_size_\u003CN>_rank_\u003CR>.pt` shard patterns, the `huggingface\u002F` metadata directory, generation config preservation, LoRA merge requirements, upload caveats, and safe preflight inspection.\n\nTypical triggers:\n\n- \"Merge an EasyR1 actor checkpoint to Hugging Face format.\"\n- \"Preflight a checkpoint before running the model merger.\"\n- \"Explain missing rank shards, unsupported DTensor placement, or LoRA base-model errors.\"\n\n## Minimal install\u002Fimport shape\n\nFor users setting up EasyR1 itself, prefer the project-documented CUDA container or an equivalent Python 3.9+ environment with compatible PyTorch, Ray, vLLM, flash-attn, transformers, and model\u002Fdataset access. A minimal import check is:\n\n```bash\npython - \u003C\u003C'PY'\nimport verl\nfrom verl.protocol import DataProto\nfrom verl.trainer.core_algos import compute_kl\nprint(\"verl\", getattr(verl, \"__version__\", \"unknown\"))\nprint(\"DataProto\", DataProto.__name__)\nprint(\"compute_kl\", compute_kl.__name__)\nPY\n```\n\nIf `flash_attn`, `vllm`, CUDA, or Ray imports fail, use the setup\u002Ftroubleshooting reference before treating EasyR1 training as ready. CPU-safe API checks do not prove the full training runtime.\n\n## Boundaries\n\nDo not use this skill as a generic TRL\u002FOpenRLHF\u002FLlamaFactory manual. Use it when EasyR1-specific config keys, scripts, dataset\u002Freward contracts, `verl` APIs, or checkpoint layout matter. For ordinary supervised fine-tuning or inference workflows that EasyR1 explicitly does not provide, route to a more suitable LLM fine-tuning or serving skill.\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\u002Feasy-r1",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 easy-r1",{"tool":40,"command":41},"bunx","bunx skills add VectorSpaceLab\u002FAREX-Skill --skill easy-r1",{"tool":43,"command":44},"pnpm","pnpm dlx skills add VectorSpaceLab\u002FAREX-Skill --skill easy-r1","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Feasy-r1\u002FSKILL.md",[47,53,57,61],{"path":48,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":51,"sourceHtmlUrl":52},"references\u002Frepo-provenance.md","REFERENCE","参考文档",2791,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Feasy-r1\u002Freferences\u002Frepo-provenance.md",{"path":54,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":55,"sourceHtmlUrl":56},"references\u002Frepo-routing-metadata.json",556,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Feasy-r1\u002Freferences\u002Frepo-routing-metadata.json",{"path":58,"resourceType":49,"resourceTypeDesc":50,"sizeBytes":59,"sourceHtmlUrl":60},"references\u002Fsetup-and-troubleshooting.md",5473,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Feasy-r1\u002Freferences\u002Fsetup-and-troubleshooting.md",{"path":62,"resourceType":63,"resourceTypeDesc":64,"sizeBytes":65,"sourceHtmlUrl":66},"scripts\u002Feasyr1_env_check.py","SCRIPT","脚本",5186,"https:\u002F\u002Fgithub.com\u002FVectorSpaceLab\u002FAREX-Skill\u002Fblob\u002Fmain\u002Fskills\u002Frepositories\u002Frepo-skills\u002Feasy-r1\u002Fscripts\u002Feasyr1_env_check.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",[],[]]