M.predict not working

Hello! I tried to follow the steps in Rossman example, in order to get prediction for another dataset. When I call m.fit(...) I get some reasonable values, but then I want to apply this model to the test set, for submission on Kaggle I use m.predict(True) (not sure if this is the right thing to do, tho) I am getting this error cuda runtime error (59) : device-side assert triggered at /opt/conda/conda-bld/pytorch_1518244421288/work/torch/lib/THC/generic/THCTensorCopy.c:70 Can someone tell me what’s going on? Thank you!