/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/device/convolution_backward_weight_specialization.hpp Source File

/home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/device/convolution_backward_weight_specialization.hpp Source File#

Composable Kernel: /home/docs/checkouts/readthedocs.org/user_builds/advanced-micro-devices-composable-kernel/checkouts/develop/include/ck/tensor_operation/gpu/device/convolution_backward_weight_specialization.hpp Source File

Go to the documentation of this file.

 // SPDX-License-Identifier: MIT
 // Copyright (c) 2018-2023, Advanced Micro Devices, Inc. All rights reserved.
  
 #pragma once
  
 #include <string>
  
 namespace ck {
 namespace tensor_operation {
 namespace device {
  
 enum struct ConvolutionBackwardWeightSpecialization
 {
     Default,
     Filter1x1Stride1Pad0,
     Filter1x1Pad0,
     OddC,
 };
  
 inline std::string
 getConvBackwardWeightSpecializationString(const ConvolutionBackwardWeightSpecialization& s)
 {
     switch(s)
     {
     case ConvolutionBackwardWeightSpecialization::Default: return "Default";
     case ConvolutionBackwardWeightSpecialization::Filter1x1Stride1Pad0:
         return "Filter1x1Stride1Pad0";
     case ConvolutionBackwardWeightSpecialization::Filter1x1Pad0: return "Filter1x1Pad0";
     case ConvolutionBackwardWeightSpecialization::OddC: return "OddC";
     default: return "Unrecognized specialization!";
     }
 }
 } // namespace device
 } // namespace tensor_operation
 } // namespace ck