TypeError: 'int' object has no attribute '__getitem__'

[renatosc]

Hi. Thank you for sharing your work.

I was trying to run your training code (python main.py --save ./logs/resnet_$RANDOM$RANDOM --depth 28 --width 10) but I keep getting the following error:

('48 ', 'group2.block0.bn0.running_var ', '(320L,) ', 'torch.cuda.FloatTensor')
('49 ', 'group2.block0.bn0.running_mean ', '(320L,) ', 'torch.cuda.FloatTensor')
('\nTotal number of parameters:', 36497146)
0%| | 0/391 [00:00<?, ?it/s]Traceback (most recent call last):
File "main.py", line 221, in
main()
File "main.py", line 217, in main
engine.train(h, train_loader, opt.epochs, optimizer)
File "/anaconda/lib/python2.7/site-packages/torchnet/engine/engine.py", line 23, in train
for sample in state['iterator']:
File "/anaconda/lib/python2.7/site-packages/tqdm/_tqdm.py", line 862, in iter
for obj in iterable:
File "/anaconda/lib/python2.7/site-packages/torch/utils/data/dataloader.py", line 174, in next
return self._process_next_batch(batch)
File "/anaconda/lib/python2.7/site-packages/torch/utils/data/dataloader.py", line 198, in _process_next_batch
raise batch.exc_type(batch.exc_msg)
TypeError: Traceback (most recent call last):
File "/anaconda/lib/python2.7/site-packages/torch/utils/data/dataloader.py", line 34, in _worker_loop
samples = collate_fn([dataset[i] for i in batch_indices])
File "/anaconda/lib/python2.7/site-packages/torchnet/dataset/transformdataset.py", line 50, in getitem
z[k] = transform(z[k])
File "/anaconda/lib/python2.7/site-packages/torchnet/transform.py", line 14, in composition
z = tr(z)
File "build/bdist.linux-x86_64/egg/torchvision/transforms.py", line 173, in call
return ImageOps.expand(img, border=self.padding, fill=self.fill)
File "/anaconda/lib/python2.7/site-packages/PIL/ImageOps.py", line 259, in expand
width = left + image.size[0] + right
TypeError: 'int' object has no attribute 'getitem'

Full log here: https://pastebin.com/E0YZNm0B

Any idea of what that could be?

I am using a Azure Data Science VM (Ubuntu) with python 2.7

[szagoruyko]

I think you need to update torchvision package.