GoodnotesIT's picture
Add dual-output Core ML speaker embedding with unchanged weights
1bb19cb verified