Wish to provide more methods and a roadmap

[qychen2001]

Thank you for your work, I really enjoy using your libraries.

However, comparing to other repositories, such as https://github.com/qile2000/LAMDA-TALENT, torch-frame supports too few methods, which will hinder its subsequent development.

I would very much like to contribute some of the code to make torch-frame support more methods, however I'm not sure which methods are already in development, as far as I know DANets and T2G-Former are already being adapted by someone.

So I hope the official torch-frame team can give a rough roadmap, while contributors can claim different tasks to avoid duplicated development.

Also, I would like to know if there is any official intention to support more methods?

[zechengz]

Hi @qychen2001, thanks in advance.
For now we are not adding additional models from our side.
So as long as the model can be formulated into the modular design with FeatureEncoder, TableConv and Decoder. Feel free to propose PRs to add them.

as far as I know DANets and T2G-Former are already being adapted by someone

Do you know who is going to add these? We can create some issues to assign and track.