DNN-for-speech-enhancement icon indicating copy to clipboard operation
DNN-for-speech-enhancement copied to clipboard

Can this program run without GPU?

Open youngsuenXMLY opened this issue 7 years ago • 2 comments

I checked the makefile for this project, and found that cuda is a dependency. I don't have a GPU, but only CPU. So is there a more general version of this project?

youngsuenXMLY avatar Aug 25 '17 08:08 youngsuenXMLY

Hi,

No, GPU is needed for this program. However, haoyu li from Tokyo university created a Tensorflow version: https://github.com/yongxuUSTC/DNN-SpeechEnhancement

Tensorflow should work well either on CPU or GPU.

Best regards, yong

yongxuUSTC avatar Aug 27 '17 09:08 yongxuUSTC

Many thanks, but there are still some problems. In DNN_sgd_ips.py from https://github.com/yongxuUSTC/DNN-SpeechEnhancement, at line 14, '/home/hyli/Data/InternData/mvn_store.mat' , I think this is a mat for mean and variance storage, please tell how to produce this mat. at line 164, '/home/hyli/Data/InternData/trainDB_lps/RawData_Part888.mat', please tell me how to get this mat or how to produce this mat. at line 227, '/home/hyli/Data/InternData/trainDB_lps_shuffle/NormContextData_Part'+str(part_num+1)+'.mat' are many mats, please tell what these mats mean and how to get or produce these mats. Your reply will be highly appreciated, thank you !

youngsuenXMLY avatar Aug 31 '17 02:08 youngsuenXMLY