HRNet/DEKR

Exhausted GPU memory with batch size of 1

wmcnally opened this issue · 2 comments

I have 12GB of GPU memory (Titan Xp) and I ran out of memory using a batch size of 1 (when running valid.py). Is this expected? Thanks.

Hi, I think the code can run on Titan Xp. You can check other settings.

Not sure what happened... It appears to be working now and it is only consuming ~3.6 GB of memory.