Open andreipit opened 2 years ago
Thank you guys, for a very interesting project! Got "RuntimeError: CUDA out of memory". Training with --cpu is very slow. Is there any other way to train on the old gpu? How to reduce the batch size?
I'm having the same issue, can you tell me how to fix it.
Thank you guys, for a very interesting project! Got "RuntimeError: CUDA out of memory". Training with --cpu is very slow. Is there any other way to train on the old gpu? How to reduce the batch size?