0

I trained a keras model inspired by the Inception Model on Tensorflow backend. The problem is, the ouput is always the same, for differents images I tested.

However, model.evaluate give me a high accuracy percentage, so, the model seems to work.

Have you an idea ? Thanks!

Pusheen_the_dev
  • 2,077
  • 4
  • 17
  • 33
  • 2
    Much more information is needed, your dataset is probably highly unbalanced. – Dr. Snoopy May 04 '17 at 07:40
  • Possible duplicate of [Why does my CIFAR 100 CNN model mainly predict two classes?](http://stackoverflow.com/questions/42705400/why-does-my-cifar-100-cnn-model-mainly-predict-two-classes) – Martin Thoma May 05 '17 at 14:38

1 Answers1

0

Finally found the answer.

I just forgot to preprocess my input for the prediction.

All is clear now!

Pusheen_the_dev
  • 2,077
  • 4
  • 17
  • 33