search

xinjli / allosaurus

Allosaurus is a pretrained universal phone recognizer for more than 2000 languages
GNU General Public License v3.0
550 stars 86 forks source link

Timestamps for phones? #24

Closed turian closed 3 years ago

turian commented 3 years ago

It would be wonderful to optionally be able to retrieve the timestamps for the phonemes. Is that possible?

[edit: I see this suggestion #20 would it be possible to add this option to the code?]

xinjli commented 3 years ago

Yeah, I also recently heard from other people requesting this feature. I am planning to add this feature this month

amand26 commented 3 years ago

Hi xinjli,

Is the time feature available now? Also can we decode the transcription?

damiankwasnyr commented 3 years ago

Hey, I was just recommended this toolkit and having timestamps would be a life saver! Any timelines? Maybe there is something I could help out with in the implementation or perhaps there is some alpha implementation already in place?

xinjli commented 3 years ago

Hi guys,

Sorry for the late update.

I just pushed the new version including the timestamp feature to both Github and pip. You can upgrade your version to enable this feature.

The current README has a section showing how to retrieve the timestamp, basically, you only need to add a new argument --timestamp=True. You can do this both in python and cmd. Just be careful that the timestamp might not be accurate in some cases.

Thanks!