Skip to content
#

preference

Here are 68 public repositories matching this topic...

AgileRL

The complete post-training stack for LLMs, plus RL for any problem. SFT, DPO and multi-turn agentic RL, with evolutionary hyperparameter optimization built in.

  • Updated Oct 15, 2026
  • Python

Add this topic to your repo

To associate your repository with the preference topic, visit your repo's landing page and select "manage topics."

Learn more