* to use {@code Predicates.<Object>in()}.
   *
   * <p>You may prefer to use a method reference (e.g., {@code target::contains}) instead of this
   * method. However, there are some subtle considerations:
   *
   * <ul>
   *   <li>The {@link Predicate} returned by this method is {@link Serializable}.
   *   <li>The {@link Predicate} returned by this method catches {@link ClassCastException} and

   * to use {@code Predicates.<Object>in()}.
   *
   * <p>You may prefer to use a method reference (e.g., {@code target::contains}) instead of this
   * method. However, there are some subtle considerations:
   *
   * <ul>
   *   <li>The {@link Predicate} returned by this method is {@link Serializable}.
   *   <li>The {@link Predicate} returned by this method catches {@link ClassCastException} and

   * includes exceptional exits, so all control paths involving signalling must be protected by a
   * finally block.
   *
   * Further optimizations of this algorithm become increasingly subtle. A wait that terminates
   * without the guard being satisfied (due to timeout, but not interrupt) can then immediately exit
   * the monitor without signalling. If it timed out without being signalled, it does not need to

    build environment of the currently installed TensorFlow package, e.g. the
    NVIDIA CUDA and NVIDIA CuDNN versions used when TensorFlow was built.

### Profiler

*   Fix a subtle use-after-free issue in `XStatVisitor::RefValue()`.

### TPU Enhancements

*   Adds 3D mesh support in TPU configurations ops.
*   Added TPU code for `FTRL` with `multiply_linear_by_lr`.