PiperOrigin-RevId: 321785542

PiperOrigin-RevId: 321431928

In response to #875

PiperOrigin-RevId: 320013295

PiperOrigin-RevId: 319311286

PiperOrigin-RevId: 317997941

Previously TFF attempted to find all the variables by inspecting the global
variable collections; however variable creation can prevent variables being
added to the global collection by explicitly setting the `collection` argument
to an empty list. This causes TFF to be unable to find the variables.

Removes now unused `graph_keys.py` module.

PiperOrigin-RevId: 317981980

PiperOrigin-RevId: 317751191

the "zipped" metrics, rather than zipping the results of many federated
aggregates.

This has been seen to significantly speed up execution, as much as 2x for 50
clients and 10 metrics.

Thanks to @jkr26 for investigation!

PiperOrigin-RevId: 314207362

Make the len(loss_fn) = len(model_outputs) check more resilient to implementation details of the keras functional api

PiperOrigin-RevId: 313886190

PiperOrigin-RevId: 313410789

Replace `import tensorflow.compat.v2 as tf` with `import tensorflow as tf` in the `learning` module.

PiperOrigin-RevId: 310572253

PiperOrigin-RevId: 310377468

PiperOrigin-RevId: 304643997

Removes hardwired training time metric from keras model. No longer used after removal of trainable model.

PiperOrigin-RevId: 304246695

PiperOrigin-RevId: 302944673

Step 1 of 3.

PiperOrigin-RevId: 297684545

PiperOrigin-RevId: 296266706

PiperOrigin-RevId: 296005970

PiperOrigin-RevId: 295827969

-   Removes `tff.learning.from_compiled_keras_model()` from API.
-   Removes unused `tff.learning._TrainableKerasModel`.
-   Removes usage of `tf.keras.Model.output_names` and
    `tf.keras.Model.loss_functions` in `tff.learning._KerasModel` to be
    compliant with upcoming Keras refactoring.

As a side-effect of this change, TFF will no longer supported (though no known
usages):

1.   Specifying multiple loss functions or loss weights via dictionaries
keyed by model output name. Instead, using ordered sequences of
functions mapping to outputs is still possible.
2.   Sublcassed `tf.keras.Model` instances cannot be used in TFF.

PiperOrigin-RevId: 295824509

Deprecated `tff.learning.from_compiled_keras_model`, point users to `tff.learning.from_keras_model`.

This is a small step into moving towards (only) using custom client training loops, rather than relying on Keras. This will also allow us to introspect Keras less, and give TFF libraries and users more flexibility and composibility.

PiperOrigin-RevId: 292603966

PiperOrigin-RevId: 291860611

PiperOrigin-RevId: 287860207

PiperOrigin-RevId: 287599572

PiperOrigin-RevId: 287587245

PiperOrigin-RevId: 286616288

PiperOrigin-RevId: 284050731

Tag keras_utils with TODO advising us of potential problem with keras models supporting masking in metrics.

PiperOrigin-RevId: 282852183

PiperOrigin-RevId: 282830397

PiperOrigin-RevId: 281347566

PiperOrigin-RevId: 280784907

PiperOrigin-RevId: 280767713

Previously tried calling shape on a nested structure of tensors, which will not
work. Instead flatten the structre and call shape on the first tensor (all
input tensors are expected to have the same first dimension size, the batch
size).

Updated example models and tests, ensured the test failed under the current
code and passes after this change.

Fixes #785

PiperOrigin-RevId: 280221692

PiperOrigin-RevId: 274896939

PiperOrigin-RevId: 264416280

Instead of checking for attribute, check that attribute value is not None.

PiperOrigin-RevId: 264409455

PiperOrigin-RevId: 263976834

Fixes #737

PiperOrigin-RevId: 260517101

PiperOrigin-RevId: 257203044

PiperOrigin-RevId: 256266909