Calculating Mean Squared Logarithmic Error in Tensorflow returns nan

Question

I\'m training an autoencoder using tensorflow, and the starter code provides a way to calculate mean squared error as the loss function.

self.mse_loss = tf.re