bugfix of groupsize dismatch with weight shape

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

bugfix of groupsize dismatch with weight shape #195

Closed WeiweiZhang1 closed 3 months ago

WeiweiZhang1 commented 3 months ago

falcon model issue fix