Jo comments

Results 5 comments of

Jo

ResourceExhaustedError (see above for traceback): OOM when allocating tensor with shape[1,12,1024,1024] and type float on /job:localhost/replica:0/task:0/device:GPU

Whats going on here? am I out of memory? I can't get it to train.

ResourceExhaustedError (see above for traceback): OOM when allocating tensor with shape[1,12,1024,1024] and type float on /job:localhost/replica:0/task:0/device:GPU

conda install -c anaconda cudnn==6.0.0 --yes seemed to fix the problem...

ResourceExhaustedError (see above for traceback): OOM when allocating tensor with shape[1,12,1024,1024] and type float on /job:localhost/replica:0/task:0/device:GPU

> @josai how did you know that you needed to do `conda install -c anaconda cudnn==6.0.0 --yes` from the error message? googling similar errors and trying their solutions until one...

ResourceExhaustedError (see above for traceback): OOM when allocating tensor with shape[1,12,1024,1024] and type float on /job:localhost/replica:0/task:0/device:GPU

> Is this with the 345M model? I've found it only just fits in a 1080TI, so anything using substantial vram like a browser running in the background can push...

Some thoughts based on the readme

Yeah it's probably a better idea to use SQL but there's never going to be more than 40 people using it on any given day so it's not a priority...