unsloth-grpo

Features and purpose

Implementation of Group Relative Policy Optimization (GRPO) for training reasoning models, optimized for 8x memory savings (triggers: GRPO, reasoning, DeepSeek-R1, reinforcement learning, RLVR, GRPOTrainer, thinking tokens).

Quick start

Follow the repository instructions for this skill and your agent. Review the files and dependencies before adding it to your workflow.