ModelCheckpoint saves best epoch as well. Added dropout option for tabnet. Adjusted RAdam for new signatures. Adjusted the training so it can take ReduceLROnPlateau. Also so that it automatically restores the best weights after training

Modified the callbacks to accomodate the neccessity of the ReduceLROnPlateau scheduler. Replace weight by pos_weight in BCEWithLogitsLoss. Added on_eval_begin method to reset metrics

Fixed a documentation error. For the tabtransformer the input_embed is a list with tuples of 2 elements, not 3

Change tab_resnet to tab_resnet_blks. Added types to tab_net.py. Cleaned up trainer by moving some methods to tab_net_utils. Adjusted tests

Added a few types, refined the docs and added tests for tabnet and the corresponding functionalities

added an option to automatically set embed size via fastai's rule of thumb. Re-structure the preprocessing module, breaking it up in a smaller modules for better debugging and 'tractability'

added explain and compute feature importance methods to the trainer. Need to add error handling and messaging since these funcionalities are intended only for tabnet. Move the general_utils module to the training module and rename it as trainer_utils. Adapted the create_explain_matrix function to WideDeep

finished implementing TabNet encoder. Next I need to implement the interpretability methods. Also adjusted WideDeep class to be able to work with TabNet. Renamed some example scripts to be consistent with each other

fixed a naming inconsistency. The Transformer encoder blocks were named 'blks'. A more consistent name relative to the other two tab models would be 'tab_transformer_blks', and so I have changed it. This leads to a small inconsistency between pypi version and that in github, which is almost irrelevant

for reasons beyond me both lates and stable doc versions as built by Read the Docs look pretty different than the html built locally, which is really frustrating. I have just built a new version pointing to master and changed the link to the documents

Tabtransformer

Added notebook on how to use custom components. Version will be still beta because I prefer to wait for V1 until I have tried it with more datasets

Added some text to the callbacks docs. Is not rendering well in tthe Read The Docs app, who knows why

After a gigantic mess with Sphinx, where, somehow, the docs looked entirely diff when built in the locally than with the app, I had to revert back to default themes. This was painful

Finished documentation. Adjusted Notebooks. Renamed WarmUp class to what it really is, FineTune. Alias the corresponding parameters so that the user can use both, finetune and alias params. Modified a few defaults

Added float type to metrics output and removed X_text from first set of callback tests, since it seems to cause some problems at testing time (I think, see what travis has to say)

Added Test. Ignored a number of typing related errors after updated to latest version of numpy (need to look into this. Ready to move on with notebooks

adjust README. Adjust 2 first notebooks and write a new one (02_2). Fixed some examples of usage. Added FullEmbeddingDropout to TabTransformer. Encapsulate alias in Classes in loss_and_obj_alias module and change some minor style text

redo all docs to match pytorch theme. Adjust all documentation to the new code structure. Remove partial imports from...everywhere. Updated examples

decoupled the model from the training process via the introduciton of a Trainer class. All seems easier now. Need to adjust all documentation. Also added a few losses and an R2 metric, as well as the possibiliy of customising virually every single component of the model

added the possibility of passing a few more loss functions and custom ones if required. Now I need to think if I want it registered as a child or not. Also, need to test all new implementation

changed the dense layer to be almost identical to that of fastai, which I really like. Changed the code accordingly. Changed the name of DeepDense and DeepDenseResnet to TabMlp and TabResnet. Change the tests acccordingly

Adjusted documentation and examples. Refactored the LRHistory callback and fix bug in the History callback. Added multiple tests and increased code coverage to 94 per cent

Mainly re-written all tests. Also, refined parameter naming and save 0 for unseen categories (as padding) for both wide and deeptabular component

first commit towards adding the TabTransformer. This will be a mix between a number of implementation (that will of course be credited in the code)