@irebai firstly, very thankful for your work.
I am using fisher corpus. the data here is in the form of .sph files, each sphere file has about 100 utterances in it. when i do spec-augmentation features are only getting generated for some utterances say 50 out of 100 and throwing an error.
here attaching is tail part in the log file
@irebai firstly, very thankful for your work. I am using fisher corpus. the data here is in the form of .sph files, each sphere file has about 100 utterances in it. when i do spec-augmentation features are only getting generated for some utterances say 50 out of 100 and throwing an error. here attaching is tail part in the log file