[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6KF07FDSDF2JZEQ085GC0F4":3,"skill-same:01M6KF07FDSDF2JZEQ085GC0F4":57},{"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":30,"linkOnly":31,"markdownContent":32,"markdownBody":33,"frontMatter":34,"refName":35,"commitSha":36,"skillDirectory":37,"forksCount":38,"repositoryHtmlUrl":39,"repositoryHomepageUrl":40,"ownerHtmlUrl":41,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":42,"install":43,"resources":56},"01M6KF07FDSDF2JZEQ085GC0F4","huggingface-trl",null,"Using Hugging Face TRL library for building custom trainers for Preference Optimization methods like DPO, SimPO, etc.","cxcscmu\u002FSkillLearnBench","cxcscmu","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F142838272?v=4","01M6KCQDFQWPDSN0MKZY96MSQP","01M6KCQDM7G70TKJ28NTSQC0RZ",92,"MIT","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fb1-one-shot-gemini-3.1-pro-preview\u002Fnlp-paper-reproduction\u002Fhuggingface-trl\u002FSKILL.md",[],[18,19,20,21,22,23,24,25,26,27,28,29],"HuggingFace","TRL","huggingface trl","library","Optimization","DPO","hugging face","methods","simpo","Preference Optimization","for","custom","2026-11-10 10:12:30",false,"---\nname: huggingface-trl\ndescription: Using Hugging Face TRL library for building custom trainers for Preference Optimization methods like DPO, SimPO, etc.\n---\n\n# Hugging Face TRL (Transformer Reinforcement Learning)\n\nTRL is a library by Hugging Face that provides tools to train language models using Reinforcement Learning and Preference Optimization (like PPO, DPO, etc.).\n\n## Custom Trainers\nTo implement custom loss functions like SimPO, you can subclass the `DPOTrainer` and override its specific methods (like `dpo_loss` or in this case, `simpo_loss`).\n\n```python\nfrom trl import DPOTrainer\nimport torch\nimport torch.nn.functional as F\n\nclass CustomTrainer(DPOTrainer):\n    def custom_loss(self, policy_chosen_logps, policy_rejected_logps):\n        pi_logratios = policy_chosen_logps - policy_rejected_logps\n        logits = pi_logratios - self.margin\n        losses = -F.logsigmoid(self.beta * logits)\n        return losses\n```\n\n## Setup & Dependencies\nCommonly requires `transformers`, `torch`, `trl`, and `accelerate`.\nMake sure you have the matching versions as defined in the project's requirement file (e.g. `environment.yml` or `requirements.txt`).\n","# Hugging Face TRL (Transformer Reinforcement Learning)\n\nTRL is a library by Hugging Face that provides tools to train language models using Reinforcement Learning and Preference Optimization (like PPO, DPO, etc.).\n\n## Custom Trainers\nTo implement custom loss functions like SimPO, you can subclass the `DPOTrainer` and override its specific methods (like `dpo_loss` or in this case, `simpo_loss`).\n\n```python\nfrom trl import DPOTrainer\nimport torch\nimport torch.nn.functional as F\n\nclass CustomTrainer(DPOTrainer):\n    def custom_loss(self, policy_chosen_logps, policy_rejected_logps):\n        pi_logratios = policy_chosen_logps - policy_rejected_logps\n        logits = pi_logratios - self.margin\n        losses = -F.logsigmoid(self.beta * logits)\n        return losses\n```\n\n## Setup & Dependencies\nCommonly requires `transformers`, `torch`, `trl`, and `accelerate`.\nMake sure you have the matching versions as defined in the project's requirement file (e.g. `environment.yml` or `requirements.txt`).\n",{"name":5,"description":7},"main","726e1f79b96b0cb3d729d0e55a1c847e0b5353ae","skills\u002Fb1-one-shot-gemini-3.1-pro-preview\u002Fnlp-paper-reproduction\u002Fhuggingface-trl",7,"https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench","https:\u002F\u002Fcxcscmu.github.io\u002FSkillLearnBench","https:\u002F\u002Fgithub.com\u002Fcxcscmu","2026-10-02 22:55:36",{"source":8,"skillName":5,"commands":44,"prompt":54,"zipAvailable":55},[45,48,51],{"tool":46,"command":47},"npx","npx skills add cxcscmu\u002FSkillLearnBench --skill huggingface-trl",{"tool":49,"command":50},"bunx","bunx skills add cxcscmu\u002FSkillLearnBench --skill huggingface-trl",{"tool":52,"command":53},"pnpm","pnpm dlx skills add cxcscmu\u002FSkillLearnBench --skill huggingface-trl","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fb1-one-shot-gemini-3.1-pro-preview\u002Fnlp-paper-reproduction\u002Fhuggingface-trl\u002FSKILL.md",true,[],[58,76,100,118,138,150],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":69,"publishedAt":30},"01M6KM5KH175P5ABP5JMX2SBDQ","object_counter","Count occurrences of an object in the image using computer vision algorithm.","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fhuman_authored\u002Fvideo-object-counting\u002Fobject_counter\u002FSKILL.md",[64],{"categoryNo":65,"categoryName":66,"taxonomyType":67,"taxonomyTypeDesc":68},"01JBSEEDCAT000000000000016","图像与多媒体","SCENARIO","使用场景",[70,71,72,73,74,75],"object counter","object","vision","computer","Computer Vision","image",{"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":85,"publishedAt":30},"01M6KM51TMVXG79MMPCS6Y8G86","travel-planning-workflow","Invoke this skill when performing travel planning. It defines a structured workflow for generating itineraries. Following this workflow is essential to ensure the final plan satisfies user requirements.","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fhuman_authored\u002Ftravel-planning\u002Ftravel-planning-workflow\u002FSKILL.md",[82],{"categoryNo":83,"categoryName":84,"taxonomyType":67,"taxonomyTypeDesc":68},"01JBSEEDCAT000000000000023","Skill与智能体开发",[86,87,88,89,90,91,92,93,94,28,95,96,97,98,99],"workflow","planning","planning workflow","travel","travel planning","travel planning workflow","structured","performing","plan","user","it","requirements","essential","itineraries",{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":30},"01M6KM4DY820XHS9YTSK5QP5HD","bundled-dataset-search-tool","invoke this skill when you need to perform database search for travel planning. This skill provides some useful pre-packaged tools to look up accommodations, attractions, cities, driving distance, flights, and restaurants from the bundled dataset.","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fhuman_authored\u002Ftravel-planning\u002Fbundled-dataset-search-tool\u002FSKILL.md",[],[107,108,109,110,111,112,113,89,114,115,90,116,28,117,87],"search","tool","dataset","bundled dataset search tool","restaurants","Tools","up","look","flights","you","database",{"skillNo":119,"name":120,"nameZh":6,"description":121,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":122,"categories":123,"keywords":124,"publishedAt":30},"01M6KM465BB9PQ0B548HQSVAWZ","glm-output","Output-processing guidance for GLM tasks. Especially useful after glm-basics and glm-calibration when you need verifier-matching metrics from output.nc and a final \u002Froot\u002Fmetrics.json.","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fhuman_authored\u002Ftemperature-simulation\u002Fglm-output\u002FSKILL.md",[],[125,126,127,28,128,129,130,131,132,133,134,135,116,136,137],"glm output","output","glm","tasks","metrics","json","basics","guidance","matching","verifier","processing","root","calibration",{"skillNo":139,"name":140,"nameZh":6,"description":141,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":142,"categories":143,"keywords":145,"publishedAt":30},"01M6KM40KY3QWDW5W97S39R2BC","glm-calibration","Calibration guidance for GLM tasks. Often most effective after glm-basics has clarified the setup; glm-output is the companion skill for exact final metric computation.","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fhuman_authored\u002Ftemperature-simulation\u002Fglm-calibration\u002FSKILL.md",[144],{"categoryNo":83,"categoryName":84,"taxonomyType":67,"taxonomyTypeDesc":68},[146,137,127,126,132,131,128,147,148,28,149],"glm calibration","setup","companion","metric",{"skillNo":151,"name":152,"nameZh":6,"description":153,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":154,"categories":155,"keywords":156,"publishedAt":30},"01M6KM3VHHVSGMGHJQB3PPT6YG","glm-basics","A good starting point for GLM calibration tasks. Use to inspect glm3.nml, confirm how GLM runs, and identify the relevant files before moving on to calibration and output evaluation.","https:\u002F\u002Fgithub.com\u002Fcxcscmu\u002FSkillLearnBench\u002Fblob\u002Fmain\u002Fskills\u002Fhuman_authored\u002Ftemperature-simulation\u002Fglm-basics\u002FSKILL.md",[],[157,127,131,158,146,137,28,126,159,160,161,162,163,164,128],"glm basics","inspect","relevant","files","good","use","how","Evaluation"]