Is the model TheBloke/vicuna-7B-1.1-HF compatible with cuda 11.7?
#7
by
anujs
- opened
I am trying to load the vicuna-7B-1.1-HF model on EC2 instance having GPU A10G, and has Build cuda_11.7.r11.7/compiler.31442593_0.
The model is loading fine, however, the model load time is high 100-120 sec. Any suggestions on the topic?