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

Consolidate dataloader&dataset_split to dataset #105

Closed wenhuach21 closed 4 months ago

wenhuach21 commented 4 months ago

@Kaihui-intel dataloader has been changed a little