TRL - Transformer Reinforcement Learning

Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.

Source: github.com/NousResearch/hermes-agent/tree/main/skills/mlops/training/trl-fine-tuning · License: MIT