trained models

[lisha1992]

Hi, thanks for sharing the codes for TabLLM. I was wondering if you provide trained models that can be used for inference?
Thanks

[stefanhgm]

Hi @lisha1992,

thanks for using TabLLM. We only used publicly available pre-trained model (T0 from Google), which we fine-tuned with t-few based on the number of examples we used for training. You can either reproducde these experiments with our code, but I would recommend to use a more up to date LLM which will probably yield a much better performance.

I hope that is helpful!

Best,
Stefan