Release 0.4.0-rc.9

0.4.0-rc.9 (2024-11-09)

Features

• add bench method & update example (#269) (45afdc7)
• add tdrzEnable option (#264) (7f495e6)
• support wav base64 for transcribe & add transcribeData for no head data (#267) (06ecd04)
• sync whisper.cpp (#266) (7b5d284)

Release 0.4.0-rc.8

0.4.0-rc.8 (2024-01-24)

Bug Fixes

• ios: sliceNSamples not reseted on prepare (ce1dae7)

Features

• sync whisper.cpp (#186) (7109406)

Release 0.4.0-rc.7

0.4.0-rc.7 (2024-01-17)

Bug Fixes

• android: use THREAD_POOL_EXECUTOR (1421c55)
• cpp: possible deref fail in vad_simple (8efd9c9)
• ios: missing job assign in transcribeFile (d1541ae)

Release 0.4.0-rc.6

0.4.0-rc.6 (2023-12-15)

Bug Fixes

• ios: copy language / initial_prompt utf8string (8204827)

Features

• add realtimeAudioMinSec option (85066fc)
• example: use base model (8d5c325)
• sync whisper.cpp (22f27e3)

Release 0.4.0-rc.5

0.4.0-rc.5 (2023-12-09)

Bug Fixes

• cpp: rename quantize / dequantize to avoid redefinition with llama.rn (3c74201)
• ios: use calloc in wsp_ggml_metal_init during ARC enabled (4a38f56)

Features

• android, example: support react-native v0.73 (#168) (ce49fce)
• android: migrate from deprecated API in jni.cpp (43a4ae2)
• android: refactor fullTranscribe jni method (#165) (ff19b49)
• cpp: unify some platform code (audio slices, utils, ...) (#166) (913954c)
• sync whisper.cpp (f1b290b) (1c0d893) (c9ec9f7)

Release 0.4.0-rc.4

0.4.0-rc.4 (2023-11-08)

Bug Fixes

• avoid last result on abort realtime transcription (58c4c96)
• rename metal resource & gguf_ defs for avoid conflit with llama.rn (d45e235)

Features

• sync whisper.cpp and enable metal on ios (#154) (aa2effb)

Release 0.4.0-rc.1

0.4.0-rc.1 (2023-10-12)

Features

• ios: make option for enable/disable Core ML (#145) (66e9a0c)

Release 0.4.0-rc.0

0.4.0-rc.0 (2023-10-11)

Features

• add new abort callback for stop transcribe (#143) (f044d6b)
• ios: add RNWHISPER_ENABLE_METAL env in podspec (34fe9a6)
• sync whisper.cpp (enable ggml-alloc) (#123) (8a58082)

Release 0.3.9

0.3.9 (2023-10-03)

Bug Fixes

• jest: add missing methods to mock (2854265)
• tvos: category DefaultToSpeaker / AllowBluetooth not available (a66e51a)

Release 0.3.8

0.3.8 (2023-10-02)

Features

• add onNewSegments prop for transcribe file (#138) (05a2648)
• ios: allow strings for coreml assets (#133) (362274f) (@JEF1056)
• ios: implement AudioSession utils & add options to transcribeRealtime (#139) (c8fc8fd)
• refactor audio utils (#140) (634132f)