Tri Dao

CUDA Templates and Python DSLs for High-Performance Linear Algebra

Toolkit for creating, sharing and using natural language prompts.

No description.

A repo for distributed training of language models with Reinforcement Learning via Human Feedback (RLHF)