s_);
  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);

  const char* tensor_data[] = {&tensor1[0], &tensor2[0]};
  const size_t tensor_size[] = {TF_ARRAYSIZE(tensor1), TF_ARRAYSIZE(tensor2)};
  const int64_t* tensor_dims[] = {&dims1[0], &dims2[0]};
  const size_t tensor_ndims[] = {ndims1, ndims2};
  for (int i = 0; i < 2; ++i) {
    TF_Tensor* v = values[i];

// of intermediate tensors since importer doesn't guarantee to preserve tensor
// names except output tensors.
Location OpLoc(const OperatorT& op,
               const std::vector<std::unique_ptr<tflite::TensorT>>& tensors,
               Builder builder, Location base) {
  if (op.outputs.empty()) return base;

  llvm::SmallVector<Location, 4> locations;
  locations.reserve(op.outputs.size());

  DataType data_type() const { return data_type_; }
  Status status() const { return status_; }
  const Tensor& tensor() const { return tensor_; }

 private:
  Status status_;
  Output output_ = Output(Operation(nullptr), 0);
  Tensor tensor_;
  const std::string node_name_ = "";
  int32 index_ = 0;
  DataType data_type_ = DT_INVALID;
};

    auto* ret = emptyTensor->tensor;
    delete emptyTensor;
    return ret;
  }

  Tensor tensor;
  if (!tensor.CopyFrom(src, src.shape())) {
    return nullptr;
  }
  return new tensorflow::TensorInterface(std::move(tensor));
}

// Non-static for testing.
TF_Tensor* TF_TensorFromTensor(const tensorflow::Tensor& src, Status* status) {

  }
}

}  // namespace

bool IsQuantized(const TensorT& tensor) {
  return (tensor.quantization != nullptr) &&
         !tensor.quantization->zero_point.empty();
}

// Returns the correct type for a quantized tensor.
// We have a special case for constants since they have a higher minimum value.
StatusOr<QuantizedType> GetQuantizedType(const TensorT& tensor, Builder builder,

// model.
table SubGraph {
  // A list of all tensors used in this subgraph.
  tensors:[Tensor];

  // Indices of the tensors that are inputs into this subgraph. Note this is
  // the list of non-static tensors that feed into the subgraph for inference.
  inputs:[int];

  // Indices of the tensors that are outputs out of this subgraph. Note this is
  // the list of output tensors that are considered the product of the

  // There should be 3 tensors: input, output, and BroadcastTo/shape.
  EXPECT_THAT(subgraph->tensors, SizeIs(3));

  // Input Tensor
  EXPECT_THAT(subgraph->tensors[0]->type, Eq(tensor_type_));
  EXPECT_THAT(subgraph->tensors[0]->name, Eq("input_1"));
  EXPECT_THAT(subgraph->tensors[0]->quantization->scale, SizeIs(1));
  EXPECT_THAT(subgraph->tensors[0]->quantization->zero_point, SizeIs(1));

// model.
table SubGraph {
  // A list of all tensors used in this subgraph.
  tensors:[Tensor];

  // Indices of the tensors that are inputs into this subgraph. Note this is
  // the list of non-static tensors that feed into the subgraph for inference.
  inputs:[int];

  // Indices of the tensors that are outputs out of this subgraph. Note this is
  // the list of output tensors that are considered the product of the

    }
  }

  // Returns whether any tensor in a list of tensors is being watched and has
  // a trainable dtype.
  bool ShouldRecord(absl::Span<const int64_t> tensor_ids,
                    absl::Span<const tensorflow::DataType> dtypes) const;

  // Adds this tensor to the list of watched tensors.
  //
  // This is a no-op if the tensor is already being watched either from an

def PromiseOp: TensorflowMlrt_Op<"promise", []> {
  let summary = "Set a tensor in a promise";

  let description = [{
    Set a tensor in a promise.

    $promise: A value of type !mlrt.promise. The underlying value must be a tensorflow tensor.
    $tensor: A tensorflow tensor.
  }];

  let arguments = (ins
    MlrtPromiseType:$promise,
    TFTensorType:$tensor
  );

  let assemblyFormat = "operands attr-dict";
}