return op_->SetAttrString(attr_name, data, length);
}
Status SetAttrInt(AbstractOperation* op_, const char* attr_name, int64_t value,
                  ForwardOperation* forward_op_) {
  forward_op_->attrs.Set(attr_name, static_cast<int64_t>(value));
  return op_->SetAttrInt(attr_name, value);
}
Status SetAttrFloat(AbstractOperation* op_, const char* attr_name, float value,

    v.reserve(num_values);
    for (int i = 0; i < num_values; ++i) {
      v.emplace_back(static_cast<const char*>(values[i]), lengths[i]);
    }
    desc->node_builder.Attr(attr_name, v);
  }
}

void TF_SetAttrInt(TF_OperationDescription* desc, const char* attr_name,
                   int64_t value) {
  desc->node_builder.Attr(attr_name, static_cast<int64_t>(value));
}

  TF_OperationDescription* reduce_desc =
      TF_NewOperation(body.get(), "CollectiveReduce", "CollectiveReduce");
  TF_SetAttrType(reduce_desc, "T", TF_FLOAT);
  TF_SetAttrInt(reduce_desc, "group_size", group_size);
  TF_SetAttrInt(reduce_desc, "group_key", 0);
  TF_SetAttrInt(reduce_desc, "instance_key", 0);

  const std::string merge_op("Mul");
  TF_SetAttrString(reduce_desc, "merge_op", merge_op.c_str(),

      0, graph, s, ::tensorflow::strings::StrCat(name, "_const0").c_str());
  TF_OperationDescription* desc = TF_NewOperation(graph, "Split", name);
  TF_AddInput(desc, {zero, 0});
  TF_AddInput(desc, {input, 0});
  TF_SetAttrInt(desc, "num_split", 3);
  TF_SetAttrType(desc, "T", TF_INT32);
  // Set device to CPU since there is no version of split for int32 on GPU
  // TODO(iga): Convert all these helpers and tests to use floats because

  forward_op_.attrs.Set(attr_name, StringPiece(data, length));
  return parent_op_->SetAttrString(attr_name, data, length);
}
Status TapeOperation::SetAttrInt(const char* attr_name, int64_t value) {
  forward_op_.attrs.Set(attr_name, static_cast<int64_t>(value));
  return parent_op_->SetAttrInt(attr_name, value);
}
Status TapeOperation::SetAttrFloat(const char* attr_name, float value) {
  forward_op_.attrs.Set(attr_name, value);

                       size_t length) override {
    tensorflow::StringPiece s(data, length);
    op_->node_builder.Attr(attr_name, s);
    return absl::OkStatus();
  }
  Status SetAttrInt(const char* attr_name, int64_t value) override {
    op_->node_builder.Attr(attr_name, static_cast<int64_t>(value));
    return absl::OkStatus();
  }
  Status SetAttrFloat(const char* attr_name, float value) override {

  if (!s.ok()) {
    LOG(WARNING) << "Unable to set attribute: " << attr_name;
  }
}

void TFE_OpSetAttrInt(TFE_Op* op, const char* attr_name, int64_t value) {
  auto s = tensorflow::unwrap(op)->SetAttrInt(attr_name, value);
  if (!s.ok()) {
    LOG(WARNING) << "Unable to set attribute: " << attr_name;
  }
}

void TFE_OpSetAttrFloat(TFE_Op* op, const char* attr_name, float value) {

    EXPECT_EQ(list[i], string(static_cast<const char*>(values[i]), lens[i]))
        << i;
  }
}

TEST_F(CApiAttributesTest, Int) {
  auto desc = init("int");
  TF_SetAttrInt(desc, "v", 31415);

  auto oper = TF_FinishOperation(desc, s_);
  ASSERT_EQ(TF_OK, TF_GetCode(s_)) << TF_Message(s_);
  EXPECT_TF_META("v", -1, TF_ATTR_INT, -1);

  int64_t value;