I couldn't get it to work well

lucidrains / tab-transformer-pytorch

Implementation of TabTransformer, attention network for tabular data, in Pytorch

MIT License

812 stars 103 forks source link

I couldn't get it to work well #4

Open delai50 opened 3 years ago

delai50 commented 3 years ago

First of all, thanks a lot for your work! Regarding this implementation, did you try it in any specific task? I was trying to apply it in a Kaggle competition with very simple tabular data (https://www.kaggle.com/c/tabular-playground-series-mar-2021) but it seems that it doesn't learn anything or doesn't work very well, and I haven't been able to find a bug yet.

patiencefromzhou1229 commented 3 years ago

Me too. All the outputs I got are 0.0, and the parameters were not updating.

pak262 commented 3 years ago

me too I get an Index error with this coding: pred = model(x_categ, x_cont) pred = model(x_categ, x_cont) -> IndexError