Implementation of the Byte-Pair Encoding Tokenizer, RoPE Embeddings, Transformer LLM distributed training & inference from scratch w/ PyTorch (and MLX), with a Flash Attention 2 Triton kernel.
Stars
0
Forks
0
Watchers
0
Open Issues
0
Overall repository health assessment
No package.json found
This might not be a Node.js project
96
commits