## Review PRs

* If a PR doesn't explain what it does or why, if it seems like it could be useful, ask for more information. Otherwise, feel free to close it.

* If a PR seems to be spam, meaningless, only to change statistics (to appear as "contributor") or similar, you can simply mark it as `invalid`, and it will be automatically closed.

   *
   * <p>The returned multimap is not thread-safe or serializable, even if the underlying multimap
   * is. The {@code equals} and {@code hashCode} methods of the returned multimap are meaningless,
   * since there is not a definition of {@code equals} or {@code hashCode} for general collections,
   * and {@code get()} will return a general {@code Collection} as opposed to a {@code List} or a
   * {@code Set}.

   *
   * <p>The returned multimap is not thread-safe or serializable, even if the underlying multimap
   * is. The {@code equals} and {@code hashCode} methods of the returned multimap are meaningless,
   * since there is not a definition of {@code equals} or {@code hashCode} for general collections,
   * and {@code get()} will return a general {@code Collection} as opposed to a {@code List} or a
   * {@code Set}.

    *   When using mixed precision, calling `RMSprop.apply_gradients` or
        `Nadam.apply_gradients` outside a `tf.function` does not work and will
        raise the AttributeError "Tensor.op is meaningless when eager execution
        is enabled". See this
        [issue](https://github.com/tensorflow/tensorflow/issues/45536) for
        details and a workaround.

## Bug Fixes and Other Changes