What's Changed
- fix(chat): clamp effort to model-supported levels by @theluckystrike in #1002
- [Feat]: One Whisper model for dictation and voice mode by @pocketpal-dev-team[bot] in #1017
- [Feat]: Faster Whisper on Android phones with dotprod by @pocketpal-dev-team[bot] in #1019
- [Bug]: Update Phonon voice add-on to fix App Store rejection by @pocketpal-dev-team[bot] in #1022
- Guard the whisper.rn dotprod loader gate by @pocketpal-dev-team[bot] in #1021
- iOS: build llama.rn from source with a precompiled Metal library by @pocketpal-dev-team[bot] in #1016
- [Bug]: iOS dictation crash when Whisper's Metal shaders fail to compile by @pocketpal-dev-team[bot] in #1020
- Translations update from Hosted Weblate by @weblate in #1018
New Contributors
- @theluckystrike made their first contribution in #1002
Full Changelog: v1.21.0...v1.21.1