Uigyu/qwen_2.5_7b-grpo_sweep_lr3

A Qwen2.5 7B-class language model checkpoint further post-trained with GRPO (Group Relative Policy Optimization) as part of a learning-rate sweep experiment. It is a community fine-tune uploaded to Hugging Face by user Uigyu, inheriting the base capabilities of the Alibaba Qwen Team's Qwen2.5-7B open-weight model. Source: supabase.