added instances for client example

library/include/ck/library/tensor_operation_instance/gpu/transpose_3d.hpp

@@ -6,7 +6,7 @@
 #include <vector>
 #include <memory>
 #include "ck/ck.hpp"
-#include "ck/tensor_operation/gpu/device/device_elementwise_3d_impl.hpp"
+#include "ck/tensor_operation/gpu/device/impl/device_elementwise_3d_impl.hpp"
 #include "ck/tensor_operation/gpu/device/tensor_layout.hpp"
 #include "ck/library/tensor_operation_instance/device_operation_instance_factory.hpp"
 
@@ -62,4 +62,4 @@ struct DeviceOperationInstanceFactory<
 } // namespace instance
 } // namespace device
 } // namespace tensor_operation
-} // namespace ck
+} // namespace ck
\ No newline at end of file

library/src/tensor_operation_instance/gpu/transpose/device_transpose_instance.cpp

+// SPDX-License-Identifier: MIT
+// Copyright (c) 2023, Advanced Micro Devices, Inc. All rights reserved.
+
+#include "ck/library/tensor_operation_instance/gpu/conv_tensor_rearrange/device_column_to_image_instance.hpp"
+#include "ck/library/tensor_operation_instance/add_device_operation_instance.hpp"
+
+namespace ck {
+namespace tensor_operation {
+namespace device {
+namespace instance {
+
+using namespace ck::transpose_op;
+
+void add_device_transpose_f16_instances(
+    std::vector<std::unique_ptr<DeviceElementwise3dImpl<F16, F16, NCDHW, 3>>>& instances)
+{
+#ifdef CK_ENABLE_FP16
+    add_device_operation_instances(instances, device_transpose_f16_instances<1, GNWC>{});
+#else
+    ignore = instances;
+#endif
+}
+
+void add_device_transpose_f32_instances(
+    std::vector<std::unique_ptr<DeviceElementwise3dImpl<F32, F32, NCDHW, 3>>>& instances)
+{
+#ifdef CK_ENABLE_FP32
+    add_device_operation_instances(instances, device_transpose_f32_instances<1, GNWC>{});
+#else
+    ignore = instances;
+#endif
+}
+
+} // namespace instance
+} // namespace device
+} // namespace tensor_operation
+} // namespace ck
\ No newline at end of file