Joint training on multiple datasets

[Shreyas-Bhat]

Dear authors,

Thank you so much for open-sourcing your code. It's very intuitive and easy to use. I am currently running experiments where I need to ingest multiple tables/datasets for joint training ( I believe the paper does this experiment). Is the current code in the repo set up to perform for this setting? if not could you please give me pointers on how I can change the current code base to do so?

Thanks for your time,
Shreyas