Quantized LLM training in pure CUDA/C++github.com2 points·ghuntley··0 commentsOpen articleSaveView on HN