I’ve developed a TensorFlow model for an artificial intelligence project, but I’m having a problem with NaN in the loss function during training. Here’s an extract from my code: