_optimizers = _KerasLazyLoader(
    globals(), submodule="optimizers", name="optimizers")
_initializers = _KerasLazyLoader(
    globals(), submodule="initializers", name="initializers")
setattr(_current_module, "losses", _losses)
setattr(_current_module, "metrics", _metrics)
setattr(_current_module, "optimizers", _optimizers)
setattr(_current_module, "initializers", _initializers)

package(
    # copybara:uncomment default_applicable_licenses = ["//tensorflow:license"],
    licenses = ["notice"],
)

# Currently pybind extension shared objects must use only C API headers since
# the C API has static initializers duplicated in the Python bindings. So we
# need a second rule that omits .cc files, in
# tensorflow/python:_pywrap_parallel_device.
filegroup(
    name = "lib_headers",
    srcs = ["parallel_device_lib.h"],
)

* Use of the alpha initializers feature now requires enabling the `Initializers` feature gate. This feature gate is automatically enabled if the `Initializers` admission plugin is enabled.([#51436](https://github.com/kubernetes/kubernetes/pull/51436))

*   RNG behavior change for `tf.keras.initializers`. Keras initializers will now
    use stateless random ops to generate random numbers.
    *   Both seeded and unseeded initializers will always generate the same
        values every time they are called (for a given variable shape). For
        unseeded initializers (`seed=None`), a random seed will be created and
        assigned at initializer creation (different initializer instances get

* Scale max-inflight limits together with master VM sizes. ([#73268](https://github.com/kubernete...

* Action required: validation rule on metadata.initializers.pending[x].name is tightened. The initializer name needs to contain at least three segments separated by dots. If you create objects with pending initializers, (i.e., not relying on apiserver adding pending initializers according to initializerconfiguration), you need to update the initializer name in existing objects and in configuration files to comply to the new validation rule. ([#512...

* [alpha] Adding admissionregistration API group which enables dynamic registration of initializers and external admission webhooks. ([#46294](https://github.com/kubernetes/kubernetes/pull/46294), [@caesarxuchao](https://github.com/caesarxuchao))


### **Application Deployment**
#### StatefulSet

// NewFromFunc allocates a new cached value instance and initializes it with an
// update function, making it ready for use.
func NewFromFunc[T any](ttl time.Duration, opts Opts, update func(ctx context.Context) (T, error)) *Cache[T] {
	return &Cache[T]{
		ttl:      ttl,
		updateFn: update,
		opts:     opts,
	}
}

// InitOnce initializes the cache with a TTL and an update function. It is
// guaranteed to be called only once.

Kubernetes with a different admission-control flag on API server. ([#58080](https://github.com/kubernetes/kubernetes/pull/58080), [@simonferquel](https://github.com/simonferquel))

* Alpha Initializers have been removed from kubeadm admission control.  Kubeadm users who still want to use Initializers can use apiServerExtraArgs through the kubeadm config file to enable it when booting up the cluster. ([#58428](https://github.com/kubernetes/kubernetes/pull/58428), [@dixudx](https://github.com/dixudx))...

- **Setup Requirements**: Requires the Istio control plane to be initialized with Pilot.

### Ambient Tests

- **Component Focus**: Primarily focuses on Ambient mode components, including `ztunnel`.
- **Setup Requirements**: Requires the Istio control plane to be initialized with `ztunnel` and other ambient components.

### Telemetry Tests