npu编译报错：error: ‘StringToDataLayout’ is not a member of ‘phi’

[shirwy]

按照https://github.com/PaddlePaddle/PaddleCustomDevice/blob/develop/backends/npu/README_cn.md
指导编译时报错
/workspace/PaddleCustomDevice/backends/npu/kernels/batch_norm_kernel.cc: In function ‘void custom_kernel::BatchNormKernel(const Context&, const phi::DenseTensor&, const phi::DenseTensor&, const phi::DenseTensor&, const paddle::optionalphi::DenseTensor&, const paddle::optionalphi::DenseTensor&, bool, float, float, const string&, bool, bool, phi::DenseTensor*, phi::DenseTensor*, phi::DenseTensor*, phi::DenseTensor*, phi::DenseTensor*, phi::DenseTensor*)’:
/workspace/PaddleCustomDevice/backends/npu/kernels/batch_norm_kernel.cc:64:33: error: ‘StringToDataLayout’ is not a member of ‘phi’
const auto data_layout = phi::StringToDataLayout(data_layout_str);
^~~~~~~~~~~~~~~~~~
/workspace/PaddleCustomDevice/backends/npu/kernels/batch_norm_kernel.cc:64:33: note: suggested alternative: ‘DataLayout’
const auto data_layout = phi::StringToDataLayout(data_layout_str);
^~~~~~~~~~~~~~~~~~
DataLayout
/workspace/PaddleCustomDevice/backends/npu/kernels/batch_norm_kernel.cc: In function ‘void custom_kernel::BatchNormGradKernel(const Context&, const phi::DenseTensor&, const paddle::optionalphi::DenseTensor&, const paddle::optionalphi::DenseTensor&, const paddle::optionalphi::DenseTensor&, const paddle::optionalphi::DenseTensor&, const phi::DenseTensor&, const phi::DenseTensor&, const paddle::optionalphi::DenseTensor&, const phi::DenseTensor&, float, float, const string&, bool, bool, bool, phi::DenseTensor*, phi::DenseTensor*, phi::DenseTensor*)’:
/workspace/PaddleCustomDevice/backends/npu/kernels/batch_norm_kernel.cc:311:33: error: ‘StringToDataLayout’ is not a member of ‘phi’
const auto data_layout = phi::StringToDataLayout(data_layout_str);
^~~~~~~~~~~~~~~~~~
/workspace/PaddleCustomDevice/backends/npu/kernels/batch_norm_kernel.cc:311:33: note: suggested alternative: ‘DataLayout’
const auto data_layout = phi::StringToDataLayout(data_layout_str);
^~~~~~~~~~~~~~~~~~
DataLayout
CMakeFiles/paddle-custom-npu.dir/build.make:276: recipe for target 'CMakeFiles/paddle-custom-npu.dir/kernels/batch_norm_kernel.cc.o' failed
make[2]: *** [CMakeFiles/paddle-custom-npu.dir/kernels/batch_norm_kernel.cc.o] Error 1
make[2]: *** Waiting for unfinished jobs....
CMakeFiles/Makefile2:109: recipe for target 'CMakeFiles/paddle-custom-npu.dir/all' failed
make[1]: *** [CMakeFiles/paddle-custom-npu.dir/all] Error 2
Makefile:102: recipe for target 'all' failed
make: *** [all] Error 2

• make_error=2
• '[' 2 '!=' 0 ']'
• echo 'Make Error Found !!!'
  Make Error Found !!!
• exit 7

[duyongtju]

我遇到了同样的问题。
CPU aarch64 架构
paddle custom divice 版本 commit ID： 15d4633
npu-smi version: 23.0.rc3

[duyongtju]

[study-hard-forever]

I had the same problem following https://github.com/PaddlePaddle/PaddleCustomDevice/blob/develop/backends/npu/README_cn.md 's process. Has it been solved？Is it the GCC and G++ compiler version? What version do you have? Mine is 7.3.0, and I see here https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/guides/hardware_support/npu_docs/paddle_install_cn.html that version 8.2+ is needed, is it related to this version? Is this the cause?

[qili93]

this issue has been fixed by PR #853, please try agian with latest code.