search

intel / auto-round

Advanced Quantization Algorithm for LLMs. This is official implementation of "Optimize Weight Rounding via Signed Gradient Descent for the Quantization of LLMs"
https://arxiv.org/abs/2309.05516
Apache License 2.0
172 stars 20 forks source link

support autoawq format #115

Closed yintong-lu closed 1 month ago

wenhuach21 commented 3 months ago

don't forget to test the pr in an env without awq lib, all the other functions should be fine

wenhuach21 commented 1 month ago

Please wait to merge until version 0.3 is released.