got error, could you give me some help?

[passion3394]

Traceback (most recent call last):
File "demo.py", line 17, in
MORAN = MORAN.cuda()
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 258, in cuda
return self._apply(lambda t: t.cuda(device))
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 185, in _apply
module._apply(fn)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 185, in _apply
module._apply(fn)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 185, in _apply
module._apply(fn)
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 112, in _apply
self.flatten_parameters()
File "/root/anaconda3/lib/python3.6/site-packages/torch/nn/modules/rnn.py", line 105, in flatten_parameters
self.batch_first, bool(self.bidirectional))
RuntimeError: CuDNN error: CUDNN_STATUS_SUCCESS

[Canjie-Luo]

There is something wrong with your CuDNN. Does it work when you run other repos?

[passion3394]

@Canjie-Luo my environment is : pytorch0.4.1 + python3.6 + cuda10.0+cudnn7.3.1, is that ok for this code?

[Canjie-Luo]

No, please ref to https://github.com/Canjie-Luo/MORAN_v2#requirements. Please use PyTorch0.3.

[passion3394]

maybe PyTorch0.3 is a little old, what's the question of the env "pytorch0.4.1 + python3.6 + cuda10.0+cudnn7.3.1" ?

[Canjie-Luo]

Sorry for that. I completed the experiments and submitted the paper one year ago. So the repo was developed under PyTorch0.2, and I upgraded it to 0.3.1 now. There seems some bugs in tools/utils.py due to the merging of Tensor and Variable in PyTorch0.4.

[passion3394]

I use another machine to run the demo.py. the env: pytorch0.4.1 + cuda9.0 + cudnn7.1. got the following error:

/home/hs/anaconda3/lib/python3.6/site-packages/torch/nn/functional.py:1961: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
"See the documentation of nn.Upsample for details.".format(mode))
Traceback (most recent call last):
File "demo.py", line 49, in
output = MORAN(image, length, text, text, test=True, debug=True)
File "/home/hs/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call
result = self.forward(*input, **kwargs)
File "/home/hs/code/MORAN_v2/models/moran.py", line 16, in forward
x_rectified, demo = self.MORN(x, test, debug=debug)
File "/home/hs/anaconda3/lib/python3.6/site-packages/torch/nn/modules/module.py", line 477, in call
result = self.forward(*input, **kwargs)
File "/home/hs/code/MORAN_v2/models/morn.py", line 109, in forward
img_offsets = (offsets_grid[i]).view(1, self.targetH, self.targetW).data.cpu().add_(-v_min).mul_(1./(v_max-v_min))
RuntimeError: Expected object of type torch.FloatTensor but found type torch.cuda.FloatTensor for argument #4 'other'

[passion3394]

@Canjie-Luo thanks for your explaining. how to fix this problem?

[Canjie-Luo]

I added an option for CUDA and updated the repo just now. Please try the new one with --cuda off.

[passion3394]

@Canjie-Luo ok,thanks very much, I will update the code and try it

[passion3394]

@Canjie-Luo I have try the new code, the code could run in cpu mode

[Canjie-Luo]

So the problem is from your CUDA environment~