Importing torchtext in Kaggle notebook

🐛 Bug

Describe the bug A clear and concise description of what the bug is.

There is a bug just by following PyTorch’s SST-2 tutorial online.

To Reproduce Steps to reproduce the behavior:

1. Go to the demo online
2. Just following the steps, code fails on import torchtext.transforms as T with torchtext==0.12.0
3. Error:

----> 1 import torchtext.transforms as T
      2 from torch.hub import load_state_dict_from_url
      3 
      4 padding_idx = 1
      5 bos_idx = 0

/opt/conda/lib/python3.7/site-packages/torchtext/transforms.py in <module>
      4 import torch
      5 from torchtext.data.functional import load_sp_model
----> 6 from torchtext.utils import get_asset_local_path
      7 from torchtext.vocab import Vocab
      8 from torchtext._torchtext import GPT2BPEEncoder as GPT2BPEEncoderPyBind, CLIPEncoder as CLIPEncoderPyBind

ImportError: cannot import name 'get_asset_local_path' from 'torchtext.utils' (/opt/conda/lib/python3.7/site-packages/torchtext/utils.py)

Expected behavior

The tutorial should work as it is.

Additional context

I’m on kaggle.com

[Nayef211]

So it looks like the default torchtext version in this tutorial is 0.11.0. I verified this using the following lines

import torchtext
torchtext.__version__

After installing torchtext 0.12.0 using !pip install torchtext==0.12.0, we are able to get past the code failure. @parmeet do you know how we can update the default torchtext and torch versions for this tutorial?

[maulberto3]

@mthrok

Nice to know the correct combination of libraries. Will be happy to report if anything else. Thanks a lot. Lastly, hopefully this will help other people too.