v0.4.5

• Fixes and new non-breaking features in useLLM and LLMModule (#406)

Full Changelog: v0.4.4...v0.4.5

v0.4.4

• Fixes multilingual Whisper not using the multilingual token (#398)

Full Changelog: v0.4.3...v0.4.4

v0.4.3

• Removed left over react-native-audio-api dependency
• Fixed error when transcribing audio file shorter than windowSize + overlapSeconds (single chunk file) with transcribe method
• Fixed being able to run inference on LLMs models when they are already generating a response.

Full Changelog: v0.4.2...v0.4.3

v0.4.2

What's Changed

• fix: use remote version of aar instead of local one by @NorbertKlockiewicz in #325

Full Changelog: v0.4.1...v0.4.2

v0.4.1

Announcing React Native ExecuTorch v0.4.1 🎉

What’s new:

🧠 More LLMs - Integrate a wide range of LLMs into your React Native app
🔧 Tool calling capabilities - Enable LLMs to dynamically interact with APIs & tools
🔍 Text Embedding Models - Transform text into vectors for semantic tasks
🗣️ Multilingual Speech to Text - accurate transcription in multiple languages
🖼️ Image Segmentation - Generate precise masks for objects in images
🌐 Multilingual OCR - Extract text from images in multiple languages

Breaking Changes:

• useLLM
• LLMModule

New Contributors

• @ArjunPatidarRadix made their first contribution in #141
• @softwarebyze made their first contribution in #143
• @patrycjakalinska made their first contribution in #284

Full Changelog: v0.3.3...v0.4.1

v0.3.3

What's Changed

• fixed bug when using Expo SDK 53

Full Changelog: v0.3.2...v0.3.3

v0.3.2

What's Changed

• feat: optimize S2T modules, removed react-native-audio-api dependency (#140)
• fix: handle local files (#150)
• docs: update README with Product Hunt badge (#146)
• add missing curly braces to example code snippet (#143)
• ci: Add Android llama demo app build check (#133)
• docs: added seo optimization (#142)
• fix: remove redundant text from speech-to-text example app (#141)
• Remove abstract pre/postprocess from BaseModel (#123)
• feat: pre-commit code formatting (#107)
• docs: Add a note of Expo Modules requirement in the docs (#132)
• fix: Make yarn v1 work with react-native-executorch (#131)

Full Changelog: v0.3.1...v0.3.2

v0.3.1

What's Changed

• fix: edit docs link by @mkopcins in #117
• chore: Bump react-native-executorch version to 0.3.0 by @jakmro in #118
• docs(readme): fix readme quickstart typo by @chmjkb in #120
• fix: fix bug in ocrs, remove node_modules from cv example app by @NorbertKlockiewicz in #122
• chore: update xcodeproj & podfiles in examples by @chmjkb in #124
• docs: add note about new architecture in readme and docs by @NorbertKlockiewicz in #127
• fix: make react-native-audio-api a peer dependency by @chmjkb in #126

Full Changelog: v0.3.0...v0.3.1

v0.3.0

We're excited to announce react-native-executorch v0.3.0. Packed with powerful new features to enhance your on-device AI experience 🎉

What's New

🗣️ Speech to Text– Transcribe your audio using Whisper and Moonshine
🎮 Speech to Text example - See how to integrate Speech to Text into your app
📖 OCR– Optical Character Recognition for extracting text from images
🚀 MPS Support – Optimized execution on iOS using Metal Performance Shaders
⚡ Hookless API – Run models outside of React components for greater flexibility
🧠 LLM Message History – Provide your LLMs with some context
⬇️ Background Model Downloads – Fetch models seamlessly in the background
📊 Benchmarks – Gain performance insights into AI models
🧪 vertical OCR - Experimental Optical Character Recognition for extracting vertical text from images

What's Changed

• docs: fix typo by @NorbertKlockiewicz in #70
• docs: correct title casing and minor typos by @chmjkb in #71
• docs: Add matching details section in the docs by @jakmro in #72
• fix: issue with isModelGenerating when switching between multiple models by @NorbertKlockiewicz in #73
• docs: Add landing page by @jakmro in #78
• docs: Improvements by @jakmro in #79
• Add video clip to README by @darthez in #85
• Update admonitions colors by @jakmro in #86
• chore: Add lefthook back by @chmjkb in #88
• fix typo tokenizer -> tokenizerSource by @mkopcins in #89
• feat: Hookless API by @jakmro in #82
• docs: Add a quickstart guide for Llama by @chmjkb in #91
• docs: Benchmarks by @jakmro in #92
• fix: Correct link in ExecuTorch bindings documentation by @jakmro in #93
• feat: Add MPS support by @jakmro in #102
• Use expo-file-system and expo-asset to fetch files by @jakmro in #97
• feat: Make third-party/ExecuTorchLib's forward() accept multiple inputs by @chmjkb in #83
• feat: Whisper by @chmjkb in #101
• fix: Remove duplicated native function in ArrayUtils.kt by @chmjkb in #106
• feat: ocr(ios) by @NorbertKlockiewicz in #84
• feat: LLM message history by @jakmro in #104
• feat: implemented vertical ocr by @NorbertKlockiewicz in #109
• design: change design of displaying ocr results by @NorbertKlockiewicz in #112
• feat: moonshine and whisper streaming by @mkopcins in #110
• docs: Add v0.3.0 documentation by @jakmro in #114
• changed isModelReady to isReady by @mkopcins in #115

New Contributors

• @darthez made their first contribution in #85

Full Changelog: v0.2.0...v0.3.0

v0.2.0

What's Changed

We are proud to announce the v0.2.0 version of React Native ExecuTorch 🚀
With this release, you can easily integrate computer vision models into your React Native apps.

Support for computer vision models 👀 :

• useObjectDetection - SSDLite320 MobileNetv3 Large
• useStyleTransfer - Candy, Mosaic, Rain Princess, Udnie
• useClassification - EfficientNet V2 Small

A hook for interacting directly with the ExecuTorch C++ Module API ⛓️ :

• useExecutorchModule

With this release, we also drop support for the React Native's old architecture, focusing on Fabric.
As always, we provide pre-exported models on our HuggingFace organization

New Contributors

• @cetfu made their first contribution in #37
• @mkopcins made their first contribution in #64

Full Changelog: v0.1.3...v0.2.0