I have my own datasets of Voxceleb1 and Voxceleb2. I am trying to find information on how to generate an .ark file so I can get access to creating an .scp file since that's how I understand it works. Is there a script for it?
Asked
Active
Viewed 43 times
0
-
There are several Python packages with support reading/writing Kaldi-compatible file formats. If your data is in the form of NumPy arrays, for example, you can check out https://github.com/csukuangfj/kaldi_native_io, which is a pybind11 wrapper around native Kaldi formats. – Desh Raj Apr 06 '23 at 16:55