dbrx icon indicating copy to clipboard operation
dbrx copied to clipboard

Loading over multiple gpus in 8bit and 4bit with transformers loader

Open RandomInternetPreson opened this issue 1 year ago • 4 comments

I can load the instruct model using the transformers loader and 8bit bits and bytes, I can get it to load evenly among multiple gpus.

However, I cannot seem to load the model with 4bit precion over multiple gpus, I managed to get the model to load across 1 24GB gpu and then start loading onto a second gpu of equivalent size, but it will not move on to any of the remaining gpus (7 in total). It will oom on the second gpu with the others sitting empty.

I've loaded other transformers based models via 4bit and never experience this heavily unbalanced loading before.

RandomInternetPreson avatar Mar 28 '24 02:03 RandomInternetPreson

how to load the model in 8bit?

jzwilliams07 avatar Mar 28 '24 06:03 jzwilliams07

I have this same issue, get an OOM/only uses a single GPU if I try to use bitsandbytes (load_in_8bit or load_in_4bit)...

lhl avatar Mar 28 '24 11:03 lhl

how to load the model in 8bit or 4bit???

huhuhu5798 avatar Apr 10 '24 08:04 huhuhu5798

Bitsandbytes library and a lot of ram

RandomInternetPreson avatar Apr 10 '24 12:04 RandomInternetPreson