search

Roche / BalancedLossNLP

Other
119 stars 23 forks source link

Versioning Problems #10

Open KemalDrop opened 1 year ago

KemalDrop commented 1 year ago

Python Version: 3.7.6 When I try to train the model with installing requirements.txt during training I am getting "packaging.version.InvalidVersion: Invalid version: '0.10.1,<0.11'". I fixed it by installing a newer version of huggingface -> 4.28.1 however at eval.py I am getting "No module named torch.fx" error with torch==1.7.0. torch.fx was released in newer versions so for inference do I need a different torch version?

blessu commented 1 year ago

Thanks for your interest. I have searched the first error message (packaging.version.InvalidVersion: Invalid version: '0.10.1,<0.11') and it turned out to be some recent issue of packaging module: https://github.com/CompVis/latent-diffusion/issues/207#issuecomment-1378716088 Please let me know if the solution suggested there does not work.

KemalDrop commented 1 year ago

The first one I solved with installing a newer version of huggingface (transformers == 4.28.1) however the second one is still an issue.

blessu commented 1 year ago

torch.fx was added in PyTorch 1.8.0, while the environment required for the experiments is transformers==4.4.2 torch==1.7.0

While you may get it through with a new pytorch version, it can be easier to solve the first one by downgrading the packaging module so that the versions of key modules stay the same.