Tabular Module

Hi - when I tried to use the tabular module (similar to Rosmann’s notebook), there was a problem with the output. Basically, when I tried to fit the learner I get a bunch of NaNs on the validation and test loss, as well well as the evaluation metric. The LR Finder graph is also empty. I tried to not use the processing method offered by Fastai (Normalize and Categorirify) and I obtained the same thing. Help?