Deal with memory constraints in training - Githubissues

aLadNamedPat / SPUR

0 stars 0 forks source link

Deal with memory constraints in training #15

Closed aLadNamedPat closed 2 weeks ago

aLadNamedPat commented 3 weeks ago

CUDA is running out of memory in training

Could be because of:

replay buffer
model size (highly unlikely)