Optimox

Hello, I have started the addition of gemma2, I will create now my PR in WIP mode. I haven't run any test yet but will do soon! Edit: My PR...

I didn't know about FlexAttention, I will look into it!

Hello @NMZ0429, Thanks for this proposal, feel free to open a PR!

There is nothing to adapt, the regression example just uses a dataset with binary targets, but if your targets are not binary then the model will learn on non binary...

feel free to open a PR with a concrete proposition.

Hello, you should have a look at this : https://www.kaggle.com/code/optimo/tabnetmultitaskclassifier If all your tasks are classification then you just need to follow this notebook and you should be good to...

Looks like one of the categories is out of range, you must have integers between 0 and cat_dim-1 for each categorical column.