curiousinspo

3 ms·

1. 1.▲
   Training a Rust 1.5B Coder LM with Reinforcement Learning (GRPO)(ghost.oxen.ai)

   3 points by curiousinspo 2y ago|0 comments
2. 2.▲
   GRPO VRAM Requirements for the GPU Poor(ghost.oxen.ai)

   3 points by curiousinspo 2y ago|1 comments
3. 3.▲
   by curiousinspo 2y ago

   Hey all, I spent some time digging into GRPO over the weekend and kicked off a bunch of fine-tuning experiments. When I saw there was already an easy to use implementation of GRPO in the trl library, I was off to the races. I broke out my l