non-experimental `Policy`, there is no loss scale associated with the `Policy`, and `Model.compile` wraps the optimizer with a `LossScaleOptimizer` if and only if the policy is a `"mixed_float16"` policy. If you previously passed a `LossScale` to the experimental `Policy`, consider just removing it, as the default loss scaling behavior is usually what you want. If you really want to customize the loss scaling behavior, you can wrap your optimizer with a `LossScaleOptimizer` before passing it to `Model.compile`....

### Avoiding permanent beta

From Kubernetes 1.20 onwards, SIG Architecture will implement a new policy to transition all REST APIs out of beta within nine months. The idea behind the new policy is to avoid features staying in beta for a long time. Once a new API enters beta, it will have nine months to either:

 - reach GA, and deprecate the beta, or

    String RESPONSE_INLINE_MIMETYPES = "response.inline.mimetypes";

    /** The key of the configuration. e.g. text/html=X-XSS-Protection: 1; mode=block<br>
     * text/html=Content-Security-Policy: reflected-xss block<br>
     * text/html=X-Frame-Options: SAMEORIGIN<br>
     *  */
    String RESPONSE_HEADERS = "response.headers";

    /** The key of the configuration. e.g. fess.search */