Acoustic model for Khalka Mongolian, trained on a corpus of 14 speakers (approx. 2 hours of speech) in Kaldi.
To use for forced alignment with the Monreal Forced Aligner, download the .zip folder and place it in the Documents/MFA/pretrained_models folder created during MFA installation. Requires a dictionary with IPA transcriptions.
Example output of alignment on training and test data in the examples folder.