Closed budach closed 6 years ago
Note: Losses still don't develop with Tensorflow 1.6
The reason for the differences seems to be that the tensorflow.keras API lags behind stand alone keras API. Not doing it for now.
@budach @fchollet what is your final solution for this important bug.
Simply replacing all occurrences of
from keras
withfrom tensorflow.python.keras
works like a charm - no crashes - BUT: the loss on my data sets doesn't improve over time anymore. Other users observe similar behavior (https://github.com/tensorflow/tensorflow/issues/15026, https://github.com/tensorflow/tensorflow/issues/15831).I will try again with a future tensorflow version. Having one less dependency is always nice.