demo.sh issues in 34b model

magic-research / PLLaVA

Official repository for the paper PLLaVA

593 stars 40 forks source link

Open Slezge opened 6 months ago

Slezge commented 6 months ago

when i run demo.sh in 34b model ,i encounter a mistake,how can i deal with it? and i already have in the meantime， i can run 13b model very well

zZer0o commented 6 months ago

you should explicitly set CUDA_VISIBLE_DEVICES like CUDA_VISIBLE_DEVICES=0,1,2,3