qwen-38-27b
🚀 Pytorch Distributed native training library for LLMs/VLMs with OOTB Hugging Face support
A C++23 library for local LLMs at the metal — inference and training, built from explicit neural-network components you can read and understand.
