/GPTQ-for-LLaMa-ROCm

4 bits quantization of LLaMA using GPTQ, ported to HIP for use in AMD GPUs.

Primary LanguagePython

Stargazers

No one’s star this repository yet.