GAN.py does not work when using tf.keras instead of the standalone keras

Question

GAN.py does not work when using tf.keras instead of the standalone keras

sfleisch opened this issue 5 years ago · comments

Get:
InvalidArgumentError: You must feed a value for placeholder tensor 'discriminator_input' with dtype float and shape [?,28,28,1]
[[{{node discriminator_input}}]]
[[{{node loss_1/model_loss/broadcast_weights/assert_broadcastable/is_valid_shape/has_valid_nonscalar_shape/has_invalid_dims/concat}}]]

I have not been able to trace this down other than it is in the generator model.

David Foster · Answer 1 · Mon Oct 14 2019 08:00:57 GMT+0800 (China Standard Time)

I have just released a new tensorflow_2 branch which maps everything across to using Keras within tensorflow 2.0. Can you see if you still get the error on this branch?

sfleisch · Answer 2 · Mon Oct 21 2019 03:42:02 GMT+0800 (China Standard Time)

The code in the tensorflow_2 branch works with the 2.0.0 release version.