Included Keras 3.3.3 upgrades for all network classes and instances of tf.model, tf.layer, and tf.optimizer
Added @tf.function decorator to coupling_networks._calculate_spline() to allow execution in non-eager environments
Changed instances of tf.Tensor in docstring to Tensor w.r.t. the experimental Tensor type
The included changes have worked successfully in all of the example / tutorial notebooks, however I'm sure there are a few introduced bugs laying around when TensorFlow is not the chosen backend.
tf.model
,tf.layer
, andtf.optimizer
@tf.function
decorator tocoupling_networks._calculate_spline()
to allow execution in non-eager environmentstf.Tensor
in docstring toTensor
w.r.t. the experimental Tensor typeThe included changes have worked successfully in all of the example / tutorial notebooks, however I'm sure there are a few introduced bugs laying around when TensorFlow is not the chosen backend.
Current environment dependencies: