News for hackers. Stuff that matters.

A comment on Mozilla Common Voice Adds 16 New Languages and 4,600 New Hours of Speech. Parent comment

Re:Mozilla Common Voice Adds 16 New Languages and 4,600 New ... by ma2rten on Thursday August 05, 2021 @01:18PM (#28076673)
Why does data for speech recognition need to be prefect. That's certainly not the case for other machine learning applications. Can you train the less clean data and fine-tune on a clean subset?
[ Parent | on HN ]
Re:Mozilla Common Voice Adds 16 New Languages and 4,600 New ... by dabinat on Thursday August 05, 2021 @01:27PM (#28076820)
Well that was kind of my point: you need to manually figure out what’s clean and what isn’t.
[ Parent | on HN ]
Re:Mozilla Common Voice Adds 16 New Languages and 4,600 New ... by ma2rten on Thursday August 05, 2021 @08:53PM (#28082027)
But it's easy to do that for a small subset for finetuning compared to cleaning up the entire dataset.
[ Parent | on HN ]