Resource exhausted: OOM when allocating tensor with shape[2] and type int32 on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc

[zhenyangGao]

When I ran bcnn_finetune.py, an error occurred: Resource exhausted: OOM when allocating tensor with shape [2] and type int32 on/job: localhost/replica: 0/task: 0/device: GPU: 0 by allocator GPU_0_bfc. It means that there is not enough memory in the gpu. I use 1080TI, 11GB RAM. Is there not enough RAM? How should I solve it? Do I need more RAM? I found a solution to reduce batch_size. Can you help me? Thank you very much.

[zhenyangGao]

I changed the input picture to 224*224, batch_size=20, and the program worked fine. If you do not change the size of the input image, my batch_size must be less than 5. It may not be enough memory, if you encounter the same problem, I hope to help you.