Run inference on multiple GPUs

#47
by itsrocchi - opened

i duplicated the space to run it on an rtx a6000, works just fine if i select 7b or 13b models but 48 gigs of vRAM are not enough for the 70B model. How can i edit the model.py script to run it on 2-3 or more GPUs? (i can add on my machine more a6000)

itsrocchi changed discussion status to closed

Sign up or log in to comment