Open ax130885 opened 1 month ago
Hi,
reduce the iteration number would lower the GPU usage but lower the performance. I have tried 9 and 11, and 9 is slightly lower.
modify the network and input size could be other options if not using pertained model.
Thank you. After changing iter to 9, I can successfully execute the training program.
glad to hear that
Hi, Sorry for replying so late.
In the ProMISe issue, it was mentioned that this project requires around 24GB of VRAM. However, I still encounter out-of-memory issue with RTX 3090 (24GB) while running the following command:
How can I resolve this problem?