IndexError: invalid index of a 0-dim tensor. Use tensor.item() to convert a 0-dim tensor to a Python number

[PanZiwei]

Hi Wangjuan,
When I tried HiCplus training process on GPU nodes, it shows an error after almost 24h. Here is the error message:

./Anaconda3/envs/plus/lib/python3.6/site-packages/torch/nn/modules/loss.py:431: UserWarning: Using a target size (torch.Size([80, 28, 28])) that is different to the input size (torch.Size([80, 1, 28, 28])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.
  return F.mse_loss(input, target, reduction=self.reduction)
0:00:00.000020
HiC version:  7
dividing, filtering and downsampling files...
(1379920, 1, 40, 40)
(1379920, 1, 40, 40)
start training...
(1379920, 1, 40, 40) (1379920, 28, 28)
Traceback (most recent call last):
  File "./Anaconda3/envs/plus/bin/hicplus", line 4, in <module>
    __import__('pkg_resources').run_script('hicplus==1.1.0', 'hicplus')
  File "./Anaconda3/envs/plus/lib/python3.6/site-packages/pkg_resources/__init__.py", line 666, in run_script
    self.require(requires)[0].run_script(script_name, ns)
  File "./envs/plus/lib/python3.6/site-packages/pkg_resources/__init__.py", line 1462, in run_script
    exec(code, namespace, namespace)
  File "./Anaconda3/envs/plus/lib/python3.6/site-packages/hicplus-1.1.0-py3.6.egg/EGG-INFO/scripts/hicplus", line 70, in <module>
    run()
  File "./Anaconda3/envs/plus/lib/python3.6/site-packages/hicplus-1.1.0-py3.6.egg/EGG-INFO/scripts/hicplus", line 66, in run
    args.func(args)
  File "./Anaconda3/envs/plus/lib/python3.6/site-packages/hicplus-1.1.0-py3.6.egg/hicplus/train_models.py", line 38, in main
    trainConvNet.train(lowres_sub,highres_sub,args.outmodel)
  File "./Anaconda3/envs/plus/lib/python3.6/site-packages/hicplus-1.1.0-py3.6.egg/hicplus/trainConvNet.py", line 120, in train
    running_loss += loss.data[0]
IndexError: invalid index of a 0-dim tensor. Use tensor.item() to convert a 0-dim tensor to a Python number

It seems that it should be a problem with PyTorch version. And here is a similar case

Do you know how to solve the problem? What PyTorch version are you using for the HiCPlus package?

Thanks!

[PanZiwei]

Finally fix the problem and my solution is below:
Change line 120 in File "./Anaconda3/envs/plus/lib/python3.6/site-packages/hicplus-1.1.0-py3.6.egg/hicplus/trainConvNet.py"

Change from
running_loss += loss.data[0]
to
running_loss += loss.data

The error reason: That's because in PyTorch>=0.5, the index of 0-dim tensor is invalid. The master branch is designed for PyTorch 0.4.1.

[wangjuan001]

Thanks for the comment! I've changed that line