Does this code offer parallel compuation? I am getting same performance on CPU and GPU?

andabi / music-source-separation

Deep neural networks for separating singing voice from music written in TensorFlow

796 stars 149 forks source link

Does this code offer parallel compuation? I am getting same performance on CPU and GPU? #22

Open bhavyaghai opened 6 years ago

bhavyaghai commented 6 years ago

Hi, I am using python3 version of this code on CentOS platform for MUSDB18 dataset. My training dataset is 100 files and test is 50 files. When I run this code on Intel Xeon processor, only 1 core is predominantly utilized out of 32. So time for each epoch is ~22min. When I switched to nVidia 1080GTX, I am still getting the same exact epoch time. Why is the performance not improving?