WapaMario63 / GPTQ-for-LLaMa-ROCm

4 bits quantization of LLaMA using GPTQ, ported to HIP for use in AMD GPUs.
32Updated last year

Alternatives and similar repositories for GPTQ-for-LLaMa-ROCm:

Users that are interested in GPTQ-for-LLaMa-ROCm are comparing it to the libraries listed below