Open polar99 opened 5 years ago
I have run src code on GTX 1080Ti with batch size 8. But I got OOM error. I have to reduce batch size to 4. How to reduce the cost of GPU memory by modifying some codes?
Tensorflow: 1.4.0 CUDA:8.0 Ubuntu: 16.04
I have run src code on GTX 1080Ti with batch size 8. But I got OOM error. I have to reduce batch size to 4. How to reduce the cost of GPU memory by modifying some codes?
Environment:
Tensorflow: 1.4.0 CUDA:8.0 Ubuntu: 16.04