GAN training and unexpected NaN results

Question

Hi everyone and happy 2021.

I am trying to train a GAN network, being the input data a dataset with 3 columns (Each with values between -1 and 1), into a NumPy array