facebookresearch/chameleon

Quantization of models?

MilanBojic1999 opened this issue · 1 comments

Hi,
I am wondering if there is any plan to make quantization of chameleon models available.
I'm working with 3060 12gb and while trying to load 7b model, I get CUDA out of memory error.

Thank you!

Unfortunately we have no plans to release a quantized model in the near future.