News for hackers. Stuff that matters.
Re:My matrix bot that transcribes speech in audio messages by jacooper on Monday December 12, 2022 @04:55PM (#33960992)
Mozilla has an open source tect-to-speech project, called deep speech.

https://github.com/mozilla/DeepSpeech

[ Parent | on HN ]
Re:My matrix bot that transcribes speech in audio messages by nukemandan on Monday December 12, 2022 @10:41PM (#33964493)
I wanted to use this, but its two tear stale with no release, seems unsupported and likely to fall into being unuseful as other models and tools exponential get better, sadly.
[ Parent | on HN ]
Re:My matrix bot that transcribes speech in audio messages by IceWreck on Tuesday December 13, 2022 @03:34AM (#33966193)
Why use google cloud's API when OpenAI released Whisper, which has 2 -5 MB models which are fast af and work so much better than any other models.

Theyre open source, run locally and require minimal computation power

[ Parent | on HN ]
Re:My matrix bot that transcribes speech in audio messages by kixiQu on Wednesday December 14, 2022 @01:16AM (#33980591)
Perhaps relevant, in response to "Why did you not share this thing":

https://twitter.com/maya_Iand/status/1441538613502570498

[ Parent | on HN ]
Re:My matrix bot that transcribes speech in audio messages by raybb on Monday December 12, 2022 @05:00PM (#33961037)
Note: this isn't my blog, just copied the title word for word
[ Parent | on HN ]