Hi, I got a weird OOM error using PyTorch 1.0.1 when I tried to run the demo.
Epoch 1/10
time: 14.569243669509888s
training loss: 0.001214
validation loss: 0.001292
test loss: 0.001461
base PSNR: 20.738712
test PSNR: 21.441355
Traceback (most recent call last):
File "main_crnn.py", line 141, in <module>
rec = rec_net(im_u, k_u, mask, test=False)
File "/home/env/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in __call__
result = self.forward(*input, **kwargs)
File "/home/env/Deep-MRI-pytorch/cascadenet_pytorch/model_pytorch.py", line 374, in forward
net['t%d_x0'%i] = self.bcrnn(x, net['t%d_x0'%(i-1)], test)
File "/home/env/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 489, in __call__
result = self.forward(*input, **kwargs)
File "/home/env/Deep-MRI-pytorch/cascadenet_pytorch/model_pytorch.py", line 301, in forward
output = output_f + output_b
RuntimeError: CUDA out of memory. Tried to allocate 60.00 MiB (GPU 0; 11.77 GiB total capacity; 8.91 GiB already allocated; 58.81 MiB free; 10.57 MiB cached)
It seems that someone else also have encountered this kind of problem and opened an issue. But it was closed by the poster.
Any help will be appreciated.
Thanks a lot.