Timothy Mak

That seems to be the problem. Thanks!

I'm getting this error now. It seems `u_name` is in `word_ctm` but not `phone_ctm`. Is that possible? I tried to debug it myself, but because of the multiprocessing, I can't...

So `sat` represents the model from the first `sat` pass and `sat_1` is the model from the second pass?

I tried re-running with the `--clean` option and this time with 2.0.0b9. I'm now getting this error. In particular, there was the `Mismatch between phone sets provided in roots file,...

FYI, my lexicon contains phones that are *case sensitive*, so phones like "T" and "t" are both present and are supposed to be different. Also, it contains phones of the...

I tried `mfa validate` and `mfa align` now as suggested. `mfa validate` did not report any error, except that it ended with a strange error message, which I don't think...

Thanks for the reply. Yes, I've been using symbolic links with different names as a work around.

It seems that breaks are just not notated in the AISHELL3 dataset. But most of the sentences are fairly short and are in fact without breaks, it seems.

If you use something like MFA (https://montreal-forced-aligner.readthedocs.io/en/latest/getting_started.html), you can probably work out where the breaks are. Or I suppose you could manually combine the shorter sentences so as to have...

But why does it insert '|' at somewhat random locations? As you can see, I have no word boundaries in `text`.