sherpa-onnx Releases

Extensive update adding SenseVoice support across multiple languages including C++, C#, Go, JavaScript, WebAssembly, Dart, and Java/Kotlin. Also introduces DirectML support and exports SenseVoice to ONNX.

Jul 23, 2024

Added a C++ runtime for MeloTTS models and support for English language. Also includes a Swift API for text punctuation and an encoder-only TRT endpointer for transducer models.

Jul 16, 2024

Introduced int8 quantized Whisper large models, exported MeloTTS to ONNX, and added pre-trained models for the Libriheavy dataset. Includes a microphone example for .NET keyword spotting.

Jul 15, 2024

Added C# keyword spotting and expanded Whisper model support to include large-v3 and distil-large-v2. Also includes a feature to find best embedding matches.

Jul 12, 2024

Jul 10, 2024

Added a Flutter text-to-speech demo and fixed iOS issues. Enhanced VAD functionality by adding a flush mechanism to detect the last audio segment.

Jul 9, 2024

This release adds platform support for iOS within the Flutter integration, expanding the mobile capabilities of the Sherpa-ONNX library.

Jul 6, 2024

This release focuses on refining the Android Flutter experience, updating build infrastructure to macOS 13, and improving ONNX Runtime configuration and library structure.

Jul 6, 2024

This release addresses stability issues, including C API crashes and ONNX Runtime integration, while adding support for .NET Framework 2.0 and the latest Silero VAD model.

Jun 29, 2024

This release introduces a real-time Flutter demo, expands Java API examples with VAD and streaming capabilities, and improves cross-platform compatibility and example structure.

Jun 25, 2024

This release focuses on new language bindings (Swift), Python enhancements (LM rescoring), and bug fixes across various components like subtitles and TTS.

Jun 22, 2024

Major update adding Dart support, Korean Android build, and Inverse Text Normalization for ASR.

Jun 18, 2024

This release includes updates to header files, fixes for WebAssembly keyword spotting, and introduces a Voice Activity Detection (VAD) example for the Dart API. It also welcomes a new contributor to the project.

Jun 14, 2024