features by binning their values
*   Improved **image preprocessing** layers: `CenterCrop`, `Rescaling`
*   Improved **image augmentation** layers: `RandomCrop`, `RandomFlip`,
    `RandomTranslation`, `RandomRotation`, `RandomHeight`, `RandomWidth`,
    `RandomZoom`, `RandomContrast`
*   Improved **`TextVectorization`** layer, which handles string tokenization,

        localAlternatingKeysAndValues = alternatingKeysAndValues;
        if (!throwIfDuplicateKeys) {
          // We want to retain only the last-put value for any given key, before sorting.
          // This could be improved, but orderEntriesByValue is rather rarely used anyway.
          localAlternatingKeysAndValues = lastEntryForEachKey(localAlternatingKeysAndValues, size);

        Entry<K, V>[] nonNullEntries = (Entry<K, V>[]) entries;
        if (!throwIfDuplicateKeys) {
          // We want to retain only the last-put value for any given key, before sorting.
          // This could be improved, but orderEntriesByValue is rather rarely used anyway.
          nonNullEntries = lastEntryForEachKey(nonNullEntries, size);
          localSize = nonNullEntries.length;
        }
        Arrays.sort(

n := 0
for i := 0; i < m; i++ {
	n += *shared
}
</pre>

into:

<pre>
n := 0
local := *shared
for i := 0; i < m; i++ {
	n += local
}
</pre>

<p>
provided it can be proved that <code>*shared</code> will not fault on access,
because the potential added read will not affect any existing concurrent reads or writes.
On the other hand, the rewrite would not be valid in a source-to-source translator.
</p>

   ([#122056](https://github.com/kubernetes/kubernetes/pull/122056), [@dhenkel92](https://github.com/dhenkel92))
- Improved scheduler performance when no scoring plugins were defined.
   ([#122058](https://github.com/kubernetes/kubernetes/pull/122058), [@aleksandra-malinowska](https://github.com/aleksandra-malinowska))
- Improved scheduler performance when no scoring plugins were defined.

access-authn-authz/admission-controllers/#noderestriction). In authentication, we added alpha-level support for automounting improved service account tokens through projected volumes. We also enabled [audience validation in TokenReview](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.13/#tokenreview-v1-authentication-k8s-io) for the new tokens for improved scoping. Under audit logging, the new alpha-level "dynamic audit configuration" adds support for [dynamically registering webhooks...

### Scheduling

- Improved fairness of the scheduling queue by placing pods which are attempted recently behind other pods with the same priority. ([#73700](https://github.com/kubernetes/kubernetes/pull/73700), [@denkensk](https://github.com/denkensk))

(kernel or iptables versions are insufficient), the kube-proxy will fall-back
to userspace mode. Kube-proxy is much more performant and less
resource-intensive in ‘iptables’ mode.
  * Node stability can be improved by reserving [resources](https://github.com/kubernetes/kubernetes/blob/release-1.2/docs/proposals/node-allocatable.md) for the base operating system using --system-reserved and --kube-reserved Kubelet flags

### Auth

- Fixed OpenID Connect (OIDC) token refresh when the client secret contains a special character. ([#76914](https://github.com/kubernetes/kubernetes/pull/76914), [@tsuna](https://github.com/tsuna))
- Improved `kubectl auth can-i` command by warning users when they try to access a resource out of scope. ([#76014](https://github.com/kubernetes/kubernetes/pull/76014), [@WanLinghao](https://github.com/WanLinghao))

- Improved memory usage of `kube-scheduler` by dropping the `.metadata.managedFields` field that `kube-scheduler` doesn't require. ([#119556](https://github.com/kubernetes/kubernetes/pull/119556), [@linxiulei](https://github.com/linxiulei))