domains / families / general / trl-training

trl-training

authentic

Post-train LLMs with TRL (Transformers Reinforcement Learning) — SFT, DPO, GRPO, KTO, and reward-model training. Use when writing or debugging training code with the TRL Python API or the trl CLI.

When to use

• SFT training
• DPO training
• GRPO training
• KTO training
• reward-model training
• Distillation training
• writing TRL Python API code
• debugging TRL training code

When not to use

• experimental trainers with unstable APIs

Skill page

Source