// Pack 3 variable handles into one TFE_TensorHandle.
  // When remote is false, function device is placed on task0. Handle types are
  // REMOTE, REMOTE, LOCAL on task0. When remote is true, function device is
  // placed on task1, Handle types are LOCAL, REMOTE, LOCAL on task1.
  int num_replicas = 3;
  std::vector<TFE_TensorHandle*> handles = {h0, h1, h2};
  TFE_TensorHandle* packed_handle =

        absl::StrCat("Trying to obtain resource handle from Input[", index,
                     "], which is not type DT_RESOURCE."));
    status->status = cc_status;
    return nullptr;
  }
  const tensorflow::ResourceHandle& handle =
      arg_tensor.flat<tensorflow::ResourceHandle>()(0);
  tensorflow::Var* variable;
  cc_status = tensorflow::LookupResource(cc_ctx, handle, &variable);

==============================================================================*/
#include "tensorflow/c/eager/gradient_checker.h"

#include <memory>

#include "absl/types/span.h"
#include "tensorflow/c/eager/abstract_tensor_handle.h"
#include "tensorflow/c/experimental/ops/math_ops.h"
#include "tensorflow/c/tf_tensor.h"

namespace tensorflow {
namespace gradients {

using namespace std;

#include "tensorflow/c/eager/gradients.h"

#include <memory>

#include "absl/container/flat_hash_set.h"
#include "absl/types/span.h"
#include "tensorflow/c/eager/abstract_context.h"
#include "tensorflow/c/eager/abstract_tensor_handle.h"
#include "tensorflow/c/eager/c_api_experimental.h"
#include "tensorflow/c/eager/c_api_test_util.h"
#include "tensorflow/c/eager/c_api_unified_experimental.h"
#include "tensorflow/c/eager/c_api_unified_experimental_internal.h"

        ->DevicePointer();
  }
  // TODO(b/175427838): It would be nice to be able to use tensorflow::isa here.
  if (!tensorflow::TensorHandle::classof(unwrapped_handle)) {
    status->status = tensorflow::errors::InvalidArgument("Invalid handle");
    return nullptr;
  }
  tensorflow::TensorHandle* handle =
      tensorflow::TensorHandleFromInterface(unwrapped_handle);

                                               int* num_handles,
                                               TF_Status* status) {
  std::vector<tensorflow::TensorHandle*> tensor_handles;
  tensor_handles.reserve(*num_handles);
  for (int i = 0; i < *num_handles; ++i) {
    tensorflow::ImmediateExecutionTensorHandle* unwrapped_handle =
        tensorflow::unwrap(handles[i]);

  if (TF_GetCode(status) != TF_OK) return nullptr;
  TFE_TensorHandle* var_handle = nullptr;
  int num_retvals = 1;
  TFE_Execute(op, &var_handle, &num_retvals, status);
  if (TF_GetCode(status) != TF_OK) return nullptr;
  TFE_DeleteOp(op);
  if (TF_GetCode(status) != TF_OK) return nullptr;
  CHECK_EQ(1, num_retvals);
  TF_DeleteStatus(status);
  return var_handle;
}

TFE_TensorHandle* CreateVariable(TFE_Context* ctx, float value,

  tracing::TracingTensorHandle* handle = nullptr;
  for (auto input : inputs) {
    PartialTensorShape shape;
    TF_RETURN_IF_ERROR(input->Shape(&shape));
    TF_RETURN_IF_ERROR(dyn_cast<tracing::TracingContext>(ctx)->AddParameter(
        input->DataType(), shape, &handle));
    params->emplace_back(handle);
  }
  return absl::OkStatus();
}

limitations under the License.
==============================================================================*/
#include "tensorflow/c/experimental/gradients/math_grad.h"

#include "tensorflow/c/eager/abstract_tensor_handle.h"
#include "tensorflow/c/eager/gradients.h"
#include "tensorflow/c/experimental/ops/array_ops.h"
#include "tensorflow/c/experimental/ops/math_ops.h"
#include "tensorflow/c/experimental/ops/nn_ops.h"

using std::vector;

      // Since the input was forwarded, we released the input handle right away
      // and hence expect the input to be forwarded to the return tensor.
      EXPECT_EQ(orig_ptr, TF_TensorData(t));
    } else {
      // In async mode we expect forwarding to work without releasing the input
      // handle since by the time the kernel is executed we have released the
      // handle in the client code.
      EXPECT_EQ(orig_ptr, TF_TensorData(t));