absl::Span<AbstractTensorHandle*> grad_inputs) override {
    // Grad for Softmax Input
    TF_RETURN_IF_ERROR(BroadcastMul(
        ctx, grad_outputs[0], forward_outputs_[1],
        grad_inputs.subspan(0, 1)));  // upstream_grad * local softmax grad

    // Grad for labels is null
    grad_inputs[1] = nullptr;
    return absl::OkStatus();
  }

    For example, to run all tests under tensorflow/python, do:

    ```bash
    bazel test ${flags} //tensorflow/python/...
    ```

    For a single component e.g. softmax op:

    ```bash
    bazel test ${flags} tensorflow/python/kernel_tests/nn_ops:softmax_op_test
    ```

    For a single/parameterized test e.g. `test_capture_variables` in
    `tensorflow/python/saved_model/load_test.py`:

the tensorflow::ops namespace to tensorflow. * Change arg order for
`{softmax,sparse_softmax,sigmoid}_cross_entropy_with_logits` to be (labels,
predictions), and force use of named args. * tf.nn.rnn_cell.* and most functions
in tf.nn.rnn.* (with the exception of dynamic_rnn and raw_rnn) are temporarily
in tf.contrib.rnn. They will be moved back into core for TF 1.2. *
`tf.nn.sampled_softmax_loss` and `tf.nn.nce_loss` have both changed their API