SimpleASR
This tutorial aims at providing some first steps for creating your own kaldi model.
I will provide different folders with different datasets. My current work includes the following:
fsdd: I have uploaded a guide on building your own digit recognition system by the freely available Free Spoken Digits Dataset.timit: Here, I used Pytorch-Kaldi with kaldi alignments. I am not 100% sure this will work.- (Work In Progress)
greek-vosk: This is the most recent directory and will probably work better than the others. It uses the already trained Greek model from alphacephei (you may also checkout the vosk documentation).
NOTE1: I assume that you have already successfully installed kaldi (if not, then check this tutorial by Josh Meyer. If you are on Windows then you may try WSL. When I created the first directories I was using WSL but some things have changed and I am not sure if everything will work.
NOTE2: I created the first two directories (fsdd, timit) a really long time ago and I have most probably made a lot of mistakes.
If something doesn't run feel free to contact me or make a PR (or open an issue and I will fix it).