What's Changed
- 💬 Add
maybe_convert_to_chatmlmap for conversational datasets by @kashif in SFT in #2862
- [SFT] fix check for AutoLigerKernelForCausalLM by @kashif in #2874
- 🍟 [SFT] Handles the dataset if it has been preprocessed by @BenasdTW in #2863
- 🧶 [GRPO][vLLM + LoRA] Move unmerge of PEFT model after weight loading by @XZ-X in #2873
- 🪂 Don't gather logits in SFT to avoid hanging by @qgallouedec in #2890
- Release: v0.15.1 by @qgallouedec
Full Changelog: https://github.com/huggingface/trl/compare/v0.15.0...v0.15.1