跳转到主要内容
rulehub

trl-training

星标19,000
分支2,242
更新时间2026年2月16日 12:45

Train and fine-tune transformer language models using TRL (Transformers Reinforcement Learning). Supports SFT, DPO, GRPO, KTO, RLOO and Reward Model training via CLI commands.

· 实时索引

安装

用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

想先保存到本地？可下载 RuleHub 当前索引到的 GitHub 仓库压缩包。

下载 Zip
文件资源管理器2 个文件
SKILL.mdreadonly
nametrl-training
descriptionTrain and fine-tune transformer language models using TRL (Transformers Reinforcement Learning). Supports SFT, DPO, GRPO...

trl-training

Train and fine-tune transformer language models using TRL (Transformers Reinforcement Learning). Supports SFT, DPO, GRPO, KTO, RLOO and Reward Model training via CLI commands.

When to use

Use this skill when the user needs help with tasks related to trl training.

Guidelines

  • Follow repository conventions in huggingface/trl
  • Prefer minimal, focused changes
  • Validate outputs before finishing