Closed Dando18 closed 5 years ago
Describe the bug Overtraining a model produces NaN loss with the cross entropy loss function.
To Reproduce Training the simple_network example several times (i.e. call fit 3 times) and the loss will become NaN and the model won't train.
Expected behavior no NaN.
Likely fixed by using numerically stable log in cross entropy.
Describe the bug Overtraining a model produces NaN loss with the cross entropy loss function.
To Reproduce Training the simple_network example several times (i.e. call fit 3 times) and the loss will become NaN and the model won't train.
Expected behavior no NaN.
Likely fixed by using numerically stable log in cross entropy.