Use iFedAvg for tabular data interoperability check

[Annie-LiGHT]

Related to #100 and #377 iFedAvg detects feature-wise biases in tabular data relative to each user.

Suggestion: use iFedAvg to test for these biases (in a "fake" training round) before entering officially into the DISCOllaboration.

• Add Interoperability check at Train your model

• Interoperability check Performs a round of FL/DL without saving the model to get the feature-wise weights/biases for that user only.

• A heat map visualisation will highlight the features in that user's data that are systematically different

• No other users biases are shared