I moved the code locally.
happy_model.fit(X_train, Y_train, epochs=10, batch_size=16,shuffle=False)
The parameter“”shuffle=False” gives the same result.
However happy_model.evaluate(X_test, Y_test) gets a different result each time.
What kind of parameter can make .evaluate have the same result?
thanks
Hey @Fatcar2002,
I ran the function multiple times in the Coursera lab, and it gave me the same results every time. So I am not sure how to reproduce the same issue. Moreover, there is no shuffling aspect involved in loading the test dataset, so my best guess is that this is due to some inherent randomization involved in your processors, which usually happens as I have observed myself while working on Kaggle GPU kernels, which you can’t possibly control with any of the parameters, but since your fit method gives the same results every time, this becomes unlikely as well. Do you mind sharing the results of different runs, while making sure that you only re-run the evaluate function, and nothing else?
Regards,
Elemento