Receiving a TypeError when executing example.sh

[tjiagoM]

I was trying to run the file example.sh. The first command runs fine, but it throws an exception on the second one:

$ python -m train --bmname=ENZYMES --assign-ratio=0.1 --hidden-dim=30 --output-dim=30 --cuda=1 --num-classes=6 --method=soft-assign

The output is as follows. Any idea what is going on? Given that I haven't changed anything, maybe the version of packages are different?

CUDA 1
Using node labels
Num training graphs:  540 ; Num validation graphs:  60
Number of graphs:  600
Number of edges:  37282
Max, avg, std of graph size:  125 , 32.46 , 14.87
Method: soft-assign
/location/diffpool/encoders.py:71: UserWarning: nn.init.xavier_uniform is now deprecated in favor of nn.init.xavier_uniform_.
  m.weight.data = init.xavier_uniform(m.weight.data, gain=nn.init.calculate_gain('relu'))
/location/diffpool/encoders.py:73: UserWarning: nn.init.constant is now deprecated in favor of nn.init.constant_.
  m.bias.data = init.constant(m.bias.data, 0.0)
/location/diffpool/encoders.py:293: UserWarning: nn.init.xavier_uniform is now deprecated in favor of nn.init.xavier_uniform_.
  m.weight.data = init.xavier_uniform(m.weight.data, gain=nn.init.calculate_gain('relu'))
/location/diffpool/encoders.py:295: UserWarning: nn.init.constant is now deprecated in favor of nn.init.constant_.
  m.bias.data = init.constant(m.bias.data, 0.0)
Epoch:  0
/location/miniconda/envs/pytorch/lib/python3.6/site-packages/torch/nn/_reduction.py:49: UserWarning: size_average and reduce args will be deprecated, please use reduction='mean' instead.
  warnings.warn(warning.format(ret))
/location/diffpool/train.py:209: UserWarning: torch.nn.utils.clip_grad_norm is now deprecated in favor of torch.nn.utils.clip_grad_norm_.
  nn.utils.clip_grad_norm(model.parameters(), args.clip)
Traceback (most recent call last):
  File "/location/miniconda/envs/pytorch/lib/python3.6/site-packages/PIL/Image.py", line 2515, in fromarray
    mode, rawmode = _fromarray_typemap[typekey]
KeyError: ((1, 1, 1800), '|u1')

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/location/miniconda/envs/pytorch/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/location/miniconda/envs/pytorch/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/location/diffpool/train.py", line 640, in <module>
    main()
  File "/location/diffpool/train.py", line 628, in main
    benchmark_task_val(prog_args, writer=writer)
  File "/location/diffpool/train.py", line 511, in benchmark_task_val
    writer=writer)
  File "/location/diffpool/train.py", line 218, in train
    log_assignment(model.assign_tensor, writer, epoch, writer_batch_idx)
  File "/location/diffpool/train.py", line 98, in log_assignment
    writer.add_image('assignment', data, epoch)
  File "/location/miniconda/envs/pytorch/lib/python3.6/site-packages/tensorboardX/writer.py", line 427, in add_image
    image(tag, img_tensor, dataformats=dataformats), global_step, walltime)
  File "/location/miniconda/envs/pytorch/lib/python3.6/site-packages/tensorboardX/summary.py", line 216, in image
    image = make_image(tensor, rescale=rescale)
  File "/location/miniconda/envs/pytorch/lib/python3.6/site-packages/tensorboardX/summary.py", line 254, in make_image
    image = Image.fromarray(tensor)
  File "/location/miniconda/envs/pytorch/lib/python3.6/site-packages/PIL/Image.py", line 2517, in fromarray
    raise TypeError("Cannot handle this data type")
TypeError: Cannot handle this data type

[chithangduong]

Try adding this before every writer.add_image() data = data.transpose(2,0,1)

Or else, you can specify the image format based on this https://github.com/lanpa/tensorboardX/blob/master/tensorboardX/writer.py#L355

[RexYing]

I think the tensorboardX interface has changed due to updates. An alternative is to use the new function: img = tensorboardX.utils.figure_to_image(fig)

It's more convenient than manually convert fig to data.