grad_outputs->push_back(NoGradient());
  // PadV2 adds a "constant_values" input.
  if (IsPadV2) {
    grad_outputs->push_back(NoGradient());
  }
  return scope.status();
}
REGISTER_GRADIENT_OP("Pad", PadGrad<false>);
REGISTER_GRADIENT_OP("PadV2", PadGrad<true>);

Status SpaceToBatchGrad(const Scope& scope, const Operation& op,

  DELEGATE = 51,
  BIDIRECTIONAL_SEQUENCE_LSTM = 52,
  CAST = 53,
  PRELU = 54,
  MAXIMUM = 55,
  ARG_MAX = 56,
  MINIMUM = 57,
  LESS = 58,
  NEG = 59,
  PADV2 = 60,
  GREATER = 61,
  GREATER_EQUAL = 62,
  LESS_EQUAL = 63,
  SELECT = 64,
  SLICE = 65,
  SIN = 66,
  TRANSPOSE_CONV = 67,
  SPARSE_TO_DENSE = 68,
  TILE = 69,
  EXPAND_DIMS = 70,

  DELEGATE = 51,
  BIDIRECTIONAL_SEQUENCE_LSTM = 52,
  CAST = 53,
  PRELU = 54,
  MAXIMUM = 55,
  ARG_MAX = 56,
  MINIMUM = 57,
  LESS = 58,
  NEG = 59,
  PADV2 = 60,
  GREATER = 61,
  GREATER_EQUAL = 62,
  LESS_EQUAL = 63,
  SELECT = 64,
  SLICE = 65,
  SIN = 66,
  TRANSPOSE_CONV = 67,
  SPARSE_TO_DENSE = 68,
  TILE = 69,
  EXPAND_DIMS = 70,

  DELEGATE = 51,
  BIDIRECTIONAL_SEQUENCE_LSTM = 52,
  CAST = 53,
  PRELU = 54,
  MAXIMUM = 55,
  ARG_MAX = 56,
  MINIMUM = 57,
  LESS = 58,
  NEG = 59,
  PADV2 = 60,
  GREATER = 61,
  GREATER_EQUAL = 62,
  LESS_EQUAL = 63,
  SELECT = 64,
  SLICE = 65,
  SIN = 66,
  TRANSPOSE_CONV = 67,
  SPARSE_TO_DENSE = 68,
  TILE = 69,
  EXPAND_DIMS = 70,

  %0 = "tfl.pseudo_const"() {value = dense<0> : tensor<1x2xi32>} : () -> tensor<1x2xi32>
  %1 = "tfl.pseudo_const"() {value = dense<0.0> : tensor<f32>} : () -> tensor<f32>
  %2 = "tfl.padv2"(%arg0, %0, %1) : (tensor<15600xf32>, tensor<1x2xi32>, tensor<f32>) -> tensor<15600xf32>
  func.return %2 : tensor<15600xf32>
  // CHECK:  return %arg0
}

// CHECK-LABEL: @ConstFoldEmbeddingLookup