I used to use Keras on top of Theano, and did performance testing of Keras on TensorFlow in parallel every TensorFlow release. The last release (0.8) of Tensor

I used to use Keras on top of Theano, and did performance testing of Keras on TensorFlow in parallel every TensorFlow release. The last release (0.8) of TensorFlow is much faster to my use case (model compilation twice as fast, execution about 10% faster) and made the switch to TensorFlow across the whole cluster about a week ago. Has been performing well and the transition was pretty seamless thanks to Keras.