GatherV2::Attrs attrs;
  attrs.batch_dims_ = 1;
  auto y = GatherV2(scope_, x, {{0}, {0}}, /*axis=*/1, attrs);
  TensorShape y_shape({2, 1});
  RunTest(x, shape, y, y_shape);
}

TEST_F(ArrayGradTest, GatherV2Grad_BatchDimWithAxis) {
  TensorShape shape({2, 1, 3});
  auto x = Placeholder(scope_, DT_FLOAT, Placeholder::Shape(shape));
  GatherV2::Attrs attrs;
  attrs.batch_dims_ = 1;

}

void TFE_OpAddAttrs(TFE_Op* op, const TFE_OpAttrs* attrs) {
  tensorflow::unwrap(op)->AddAttrs(tensorflow::unwrap(attrs));
}

void TFE_OpAttrsSerialize(const TFE_OpAttrs* attrs, TF_Buffer* buf,
                          TF_Status* status) {
  tensorflow::NameAttrList name_and_attrs;
  tensorflow::unwrap(attrs)->GetNameAttrList(&name_and_attrs);
  status->status = MessageToBuffer(name_and_attrs, buf);

    auto metadata_op = llvm::dyn_cast<mlir::TF::TPUReplicateMetadataOp>(op);
    if (!metadata_op) continue;

    NamedAttrList attrs(metadata_op->getAttrDictionary());

    // Missing or bad `_replication_info` attribute.
    auto replication_info_attr = attrs.get(mlir::TF::kReplicationInfoAttr);
    if (!replication_info_attr)
      return metadata_op.emitError() << kBadReplicateInfoAttrMsg;

  input: "MobilenetV1/Conv2d_0/weights/read"
  attr {
    key: "T"
    value {
      type: DT_FLOAT
    }
  }
  attr {
    key: "data_format"
    value {
      s: "NDHWC"
    }
  }
  attr {
    key: "dilations"
    value {
      list {
        i: 1
        i: 1
        i: 1
        i: 1
      }
    }
  }
  attr {
    key: "padding"
    value {

        buildFile << """
            apply plugin: 'ivy-publish'

            group = 'group'
            version = '1.0'

            def comp = new TestComponent()
            def attr1 = Attribute.of('custom', String)
            def attr2 = Attribute.of('nice', Boolean)

            comp.usages.add(new TestUsage(
                    name: 'api',
                    usage: objects.named(Usage, 'api'),

                  .Input(inputs)
                  .Attr("ids", ids)
                  .Attr("calibration_methods", calibration_methods)
                  .Finalize(node_def()));
  ASSERT_THAT(InitOp(),
              StatusIs(tsl::error::INVALID_ARGUMENT,
                       HasSubstr("NodeDef missing attr 'output_file_path'")));
}

TEST_F(CalibrationStatisticsSaverTest, WrongNumInputs) {

  }
  // TODO(b/166669239): This is needed to support AttrBuilder::Get for string
  // attributes. Number type attrs and DataType attrs work fine without this.
  // Consider getting rid of this and making the behavior between number types
  // and string consistent.
  forward_op_->attrs.BuildNodeDef();
  std::unique_ptr<GradientFunction> gradient_fn;
  TF_RETURN_IF_ERROR(registry.Lookup(*forward_op_, &gradient_fn));

	interfaces := &admission.RuntimeObjectInterfaces{EquivalentResourceMapper: mapper}

	testcases := []struct {
		name string

		webhook *v1.ValidatingWebhook
		attrs   admission.Attributes

		expectCall            bool
		expectErr             string
		expectCallResource    schema.GroupVersionResource
		expectCallSubresource string
		expectCallKind        schema.GroupVersionKind

  %filter = tfr.constant [1, 2, 2, 1] -> !tfr.attr
  %padding = tfr.constant "VALID" -> !tfr.attr
  %explicit_paddings = tfr.constant [] -> !tfr.attr
  %data_format = tfr.constant "NHWC" -> !tfr.attr
  %MaxPool = tfr.call @tf__max_pool(%input_, %stride, %filter, %padding, %explicit_paddings, %data_format) : (!tfr.tensor, !tfr.attr, !tfr.attr, !tfr.attr, !tfr.attr, !tfr.attr) -> (!tfr.tensor)

        jar.manifest.mainAttributes.getValue('attr') == null

        when: "Attribute added"
        buildFile.text = jarWithManifest("attributes(attr: 'Hello')")
        run 'jar'

        then:
        executedAndNotSkipped ':jar'
        jar.manifest.mainAttributes.getValue('attr') == 'Hello'

        when: "Attribute modified"
        buildFile.text = jarWithManifest("attributes(attr: 'Hi')")
        run 'jar'