A andrewkchan.dev Fast LLM Inference From Scratch (using CUDA) 344 points homarp 1 year ago 57 comments
