idiap / pkwrap

A pytorch wrapper for LF-MMI training and parallel training in Kaldi
Other
72 stars 12 forks source link

error: ‘cudaDataType_t’ does not name a type #17

Closed MG623 closed 2 years ago

MG623 commented 2 years ago

Hello,during "make" step of pkwrap-master, an error occured. Below is the log, can you figure out what's the problem?

[log] cc1plus: warning: command line option ‘-Wstrict-prototypes’ is valid for C/ObjC but not for C++ In file included from /home/hcg/kaldi-master/src/cudamatrix/cu-value.h:25:0, from /home/hcg/kaldi-master/src/cudamatrix/cu-vector.h:33, from /home/hcg/kaldi-master/src/cudamatrix/cu-matrix-lib.h:25, from /home/hcg/kaldi-master/src/nnet3/nnet-common.h:28, from /home/hcg/kaldi-master/src/nnet3/nnet-nnet.h:26, from /home/hcg/kaldi-master/src/nnet3/am-nnet-simple.h:25, from src/common.h:16, from src/pkwrap-main.h:6, from src/pkwrap-main.cc:4: /home/hcg/kaldi-master/src/cudamatrix/cu-device.h:112:3: error: ‘cudaDataType_t’ does not name a type cudaDataType_t GetCublasComputeType() const { return cublas_computetype; } ^ /home/hcg/kaldi-master/src/cudamatrix/cu-device.h:114:3: error: ‘cublasGemmAlgo_t’ does not name a type cublasGemmAlgo_t GetCublasGemmAlgo() const { return cublas_gemmalgo; } ^ /home/hcg/kaldi-master/src/cudamatrix/cu-device.h:381:3: error: ‘cudaDataType_t’ does not name a type cudaDataType_t cublas_computetype; ^ /home/hcg/kaldi-master/src/cudamatrix/cu-device.h:383:3: error: ‘cublasGemmAlgo_t’ does not name a type cublasGemmAlgo_t cublas_gemmalgo; ^ error: command 'gcc' failed with exit status 1 Makefile:4: recipe for target 'package' failed make: *** [package] Error 1

The environment is as follows ubuntu1~16.04.12 gcc version 5.4.0 torch 1.8.1+cu111

mrsrikanth commented 2 years ago

Hello,

Could you please tell me which version of Kaldi you are using? It will help me reproduce the error.

MG623 commented 2 years ago

Thank you very much for your reminder. I tried to change the version of Kaldi (from the latest version "kaldi-master" to version "kaldi-trunk" in January 2020). Now it has been compiled successfully.Thanks again