Missing alignments

[Imatgay]

Hi everyone! I’m working on a large-scale alignment using pretrained models and noticed that not all audio files are consistently exported as TextGrid files. For instance, during validation, the debugger logs something like: “Aligned 1796, errors on 35, total 1831”.

I’m curious if there’s a way to pinpoint the exact reasons for these alignment errors. Also, would using something like mfa adapt potentially help improve the alignment results? Any advice or insights would be greatly appreciated! Thank you in advance!”