You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm trying to use sharding for performing data parallel training. When I use only one GPU with batchsize 4, it works. But when I use a global batchsize equals 8, and sharding the data to 2 GPUs, which means batchsize=4 / GPU, I get a OOM. Why this gonna happens?
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi guys:
I'm trying to use sharding for performing data parallel training. When I use only one GPU with batchsize 4, it works. But when I use a global batchsize equals 8, and sharding the data to 2 GPUs, which means batchsize=4 / GPU, I get a OOM. Why this gonna happens?
Beta Was this translation helpful? Give feedback.
All reactions