New Dataset and OCR variational pipeline

[J-Dymond]

This branch includes:

• src/arc_spice/data/multieurlex_utils.py

  • Here code for loading and preprocessing the MultiEURLEX dataset is located.

• src/arc_spice/variational_pipelines/RTC_variational_pipeline.py

  • Here the variational pipeline is located. It has clean_inference and variational_inference functionality. As well as calculating some confidence metrics on the outputs.

• src/arc_spice/variational_pipelines/dropout_utils.py

  • This file contains some utility functions for performing MC dropout.

• src/arc_spice/eval/classification_error.py and src/arc_spice/eval/translation_error.py

  • These contain some helper functions for calculating errors and uncertainties relating to the two tasks

• scripts/variational_RTC_example.py

  • This is a barebones script with example usage of the variational pipeline.

[lannelin]

leaving a partial review before I jump into meetings.

Overall looks good, I've added some comments requesting some small changes.

[lannelin]

outstanding TODO - I think a simple fix in the data loading, will comment separately.

[lannelin]

is this [0] relying on the fact that text is a str and never a list of strings? if so, maybe add a guard to check it is a str

[J-Dymond]

Yes, I think this assumes a batch size of 1

[lannelin]

Functionality looks good.

I resolved the merge conflict for .gitignore as it was stopping the CI checks from running. Those are failing at the moment so could you take a look at why @J-Dymond ? I suspect you haven't got pre-commit installed locally when you're committing up. If you haven't, try:

pip install -e ".[dev]" # from project dir, installs with dev deps
pre-commit install
pre-commit run --all-files

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[lannelin]

nice work! A couple of very minor comments to check but then should be good to merge.