Environment:

• OS: Ubuntu 16.04

• PyTorch version: 0.3.1

• How you installed PyTorch (conda, pip, source): conda

• Python version: 3.6

• CUDA/cuDNN version: 9.1

• GPU models and configuration: Nvidia GTX 1080

• GCC version (if compiling from source): 4.8.5

• A script to reproduce the bug. Please try to provide as minimal of a test case as possible.
  Since it's dataset related, I can provide the code that loads the data directory.
  

• Error messages and/or stack traces of the bug
  

• Context around what you are trying to do
  I'm trying to train a image classifier with the official densenet201 model, the dataloader caused this error is the train data loader. It occurs randomly at different epoch of the training process, which is weird.

anyone has ideas how to solve it? please help, thanks!