Ahmadzei's picture
added 3 more tables for large emb model
5fa1a76
Transformers supports the AWQ and GPTQ quantization algorithms and it supports 8-bit and 4-bit quantization with bitsandbytes.