fastLLaMa: An experimental high-performance framework for running Decoder-only LLMs with 4-bit quantization in Python using a C/C++ backend.
Primary LanguageCMIT LicenseMIT
No one’s star this repository yet.