TFT-Interpretability

[YA2424]

Added Interpretability for Temporal Fusion Transformer.

Changed the TFT class to return the needed parameters for interpretability and implemented methods for interpretability plots.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[CLAassistant]

All committers have signed the CLA.

[marcopeix]

Thanks for the contribution! Here are a few comments to improve your PR and get it merged:

• Merge the main branch into your working branch to update it
• There a lot of changes made to documentation notebooks that shouldn't be modified. Please, revert those changes so they match what is on the main branch.
• Ideally, we don't introduce new dependencies. Can you make it such that there's a method or attribute to extract the parameters, but then it's up to the user to download any plotting library to interpret them and plot them.

Let me know if you have any questions!

[YA2424]

Thanks for the contribution! Here are a few comments to improve your PR and get it merged:

• Merge the main branch into your working branch to update it
• There a lot of changes made to documentation notebooks that shouldn't be modified. Please, revert those changes so they match what is on the main branch.
• Ideally, we don't introduce new dependencies. Can you make it such that there's a method or attribute to extract the parameters, but then it's up to the user to download any plotting library to interpret them and plot them.

Let me know if you have any questions!

Thanks for your response and comments.

All the changes you asked for are done.
Please, let me know if other changes are needed.

[marcopeix]

@amirouyanis Thanks again for the PR, great work! Can you just add this header | eval: false to all the cells where you actually run the model? So all the cells after "Usage Example". That will make the tests pass. Thank you!

[YA2424]

@amirouyanis Thanks again for the PR, great work! Can you just add this header | eval: false to all the cells where you actually run the model? So all the cells after "Usage Example". That will make the tests pass. Thank you!

@marcopeix Thank you :). The chages are Done!