Omni SenseVoice: High-Speed Speech Recognition with Words Timestamps

Re:Omni SenseVoice: High-Speed Speech Recognition with Words... by Etheryte on Sunday October 13, 2024 @05:36AM (#41826506)

This uses SenseVoice under the hood, which claims to have better accuracy than Whisper. Not sure how accurate that statement is though, since I haven't seen a third party comparison, in this space it's very easy to toot your own horn.

[0] https://github.com/FunAudioLLM/SenseVoice

[ Parent | on HN ]

Re:Omni SenseVoice: High-Speed Speech Recognition with Words... by pferdone on Sunday October 13, 2024 @07:04AM (#41826943)

I mean they make a bold statement up top just to paddle back a little bit further down with: "[…] In terms of Chinese and Cantonese recognition, the SenseVoice-Small model has advantages."

It feels dishonest to me.

[0] https://github.com/FunAudioLLM/SenseVoice?tab=readme-ov-file...

[ Parent | on HN ]

Re:Omni SenseVoice: High-Speed Speech Recognition with Words... by jmward01 on Sunday October 13, 2024 @10:45AM (#41828467)

This uses SenseVoice small under the hood. They claim their large model is better than Whisper large v3, not the small version. This small version is definitely worse than Whisper large v3 but still usable and the extra annotation it does is interesting.

[ Parent | on HN ]

Re:Omni SenseVoice: High-Speed Speech Recognition with Words... by khimaros on Sunday October 13, 2024 @04:31PM (#41831323)

this claims to have speaker diarization which is a potentially killer feature missing from most whisper implementations.

[ Parent | on HN ]