`.predict` is available for Cloud TPUs, Cloud TPU, for all types of
        Keras models (sequential, functional and subclassing models).
    *   Automatic outside compilation is now enabled for Cloud TPUs. This allows
        `tf.summary` to be used more conveniently with Cloud TPUs.
    *   Dynamic batch sizes with DistributionStrategy and Keras are supported on
        Cloud TPUs.

  let summary = "A TPU core selector Op.";

  let description = [{
This Op produces a set of TPU cores (for warm-up) or a single TPU core
(for regular inference) to execute the TPU program on. The output is
consumed by TPUPartitionedCall.
  }];

  let arguments = (ins);

  let results = (outs
    Res<TF_Int32Tensor, [{A vector 1 or more TPU cores.}]>:$device_ordinals
  );
}

\300\227\305\321\231\314\240?\215\325\361{\004\t\355?\264\200I\230\n\266\351?\340f\207\200\027\\\353?`\236\276\215\022\320\325?\254v\033H\212\320\321? H\316&1\\\267?9\314\220\201C\246\357?\3407vX]\222\232?\304G\020\241e#\344?\302\227&\237\261f\353?PU0\375\007\343\246?la\237YU\020\356?3\201\337N\200\320\356?\206\224\306gy\334\335?\020R\334\313\233\006\321?\327\006x8X*\347?\375\372\315\205\235N\340?\330\260\006\357\356f\353?sT\265\367\332\010\340?\202\366s\"Z=\326?\322AyH4M\337?\313J\374\271\256\3...