Log_softmax function returns “nan” in raspbian OS running torch wheels, but works fine in anaconda (windows OS)

Question

I have been working on a Pytorch model for one of my testing purposes. So far I trained the model using anaconda in windows and the model is working fine. I wanted to train