Added scripts on how to use the library for recsys in response to issue #133. Also Added a simple/basic transformer model for the text component before integrating with HF. Also added the option of specify the dimension of the feed forward network

first step towards adding an example on how to reproduce a kaggle notebook (details in the code) with this library

Feature importance

Added a notebook to illustrate the use of feature importance. And also fixed a little bug related to the calling of the callbacks.

Makint it consistent with master so I don't have to rebase or anything similar. Also fixing bugs related to the other forms of training where I did not fix the restore weights functionality

Fix #175 early stopping and model checkpoint restoring weights.

Fix #175 early stopping and model checkpoint restoring weights. Also fix #174 the finetunner single batch run

revert

Fix: EarlyStopping does not store and restore the model #175

manual trigger on fork PR

Get fork PR tests running

Unit tested feature importance and made it compatible to the #171 fix. I just want to add a notebook and rready to merge/publish

Unit tested feature importance and made it compatible to the #171 fix. I just want to add a notebook and rready to merge/publish

Default vision models

Unit tested the feature importance functionality for all models but Tabnet. Only TabNet left to do and then ready to publish

added oversample_mul to the dataloader_params to fix issue #153

Added a class to compute feature importance. I need to extend it to an explain capability and test the hell out of it

Cont cols quantization