rocky/OSGKeyboard
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: migrate on-device Qwen3 ASR to CoreML for background Flow dictation

Browse Source 
Replace MLX GPU inference with CoreML bundles so transcription continues
while the host app is backgrounded. Adds model download and warm-up,
vendored Qwen3Speech, and updates onboarding, settings, and copy for the
~1.6 GB CoreML package (iOS 18+).
This commit is contained in:
Rocky
2026-06-23 00:46:58 +08:00
parent 5e5122f172
commit df1c5ff32c
160 changed files with 22080 additions and 492 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+20 -6
View File
@@ -7,13 +7,27 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
## [Unreleased]
### Changed
- **iPhone only**: Set `TARGETED_DEVICE_FAMILY` to `"1"` for both `OSGKeyboard` and `OSGKeyboardExt` targets. Removed `UIRequiresFullScreen` and trimmed `UISupportedInterfaceOrientations` to Portrait only (iPad is no longer a supported device).
- **Remove top divider line**: Deleted the 0.5 pt `palette.divider` overlay from `KeyboardRootView` — the subtle highlight gradient is retained; the hard separator line is gone.
- **Keyboard preview always dark**: `KeyboardPreviewSheet` now injects `.environment(\.themePalette, Palette.dark)` alongside `.environment(\.colorScheme, .dark)` on `KeyboardPreviewStub`, so the preview palette is always the dark variant regardless of the app's active theme.
- **Docs consistency**: README/README.zh now consistently describe the currently implemented capability set (`iOS 26+`, on-device `SpeechAnalyzer` + `DictationTranscriber`) with no deferred-ASR wording.
## [0.2.0] - 2026-06-22
## [0.1.2] - In Progress
### Added
- **Local engine with on-device Qwen models**: Optional Qwen3-ASR 0.6B speech recognition and Qwen3.5-0.8B text polish, fully offline after download.
- **On-device model management**: Download, progress, delete, and readiness status in Settings; mirror auto-selection between ModelScope and Hugging Face with fallback.
- **Engine picker**: Choose between local (on-device ASR + polish) and cloud (ASR + user-configured LLM polish).
- **Flow session dictation**: TypeWhisper-style continuous capture in the host app with keyboard handoff via App Group.
- **Open-source licenses** screen for bundled third-party components.
### Changed
- **Settings simplified**: Merged language and model sections; cloud mode always enables polish (removed off/transcribe mode picker).
- **Keyboard UI**: Local/cloud engine badges replace the mode menu; shows model-not-downloaded guidance when the local stack is incomplete.
- **iPhone only**: Set `TARGETED_DEVICE_FAMILY` to `"1"` for both targets; portrait-only orientations.
- **Keyboard preview always dark**: Preview injects the dark palette regardless of app theme.
- **Docs consistency**: README/README.zh aligned to the iOS 26+ capability set.
### Fixed
- **ModelScope download progress**: Progress now tracks byte counts instead of jumping to 50% after the first file.
- **Light/Dark mode consistency** for shared button/card modifiers via `@Environment(\.themePalette)`.
## [0.1.2] - 2026-06-20
### Fixed
- **Light/Dark mode consistency**: `cardSurface()`, `primaryButton()`, `secondaryButton()`, and `pillChip()` view modifiers in `Theme.swift` now use `ViewModifier` structs that read from `@Environment(\.themePalette)`. Previously they used hardcoded dark `Palette` constants, causing cards and buttons to always render in dark mode even when the main App was in light mode.