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
245 stars 20 forks source link

add runable script for autoround #225

Closed n1ck-guo closed 1 month ago

n1ck-guo commented 2 months ago

@wenhuach21 ready to merge

wenhuach21 commented 2 months ago

Is there a way to test this with unit tests, for example, you forget to add lm-eval in the requirement.

wenhuach21 commented 2 months ago

Please ensure that all parameters are synchronized with our API arguments to avoid any confusion.