Release v1.9.10

What's Changed

• Fix CI tests for Python and JNI. by @csukuangfj in #554
• Add a new Persian tts model by @csukuangfj in #555
• Add TTS demo for C# API by @csukuangfj in #557

Full Changelog: v1.9.9...v1.9.10

Release v1.9.9

What's Changed

• Fix kws ci by @pkufool in #540
• Fix cmake variables to point to the project root directory. by @csukuangfj in #545
• add blank_penalty for offline transducer by @chiiyeh in #542
• add hotwords docstring to offline_recognizer and online_recognizer by @chiiyeh in #546
• add blank_penalty for online transducer by @chiiyeh in #548
• Fixes issue #535 , fix hexa 1-char tokens in ASR output. by @vesis84 in #550
• Ensure input for speaker ID is a valid number. by @csukuangfj in #552
• Run TTS engine service without starting the app. by @csukuangfj in #553

New Contributors

• @chiiyeh made their first contribution in #542
• @vesis84 made their first contribution in #550

Full Changelog: v1.9.8...v1.9.9

Release v1.9.8

What's Changed

• Add missing field for two-pass APK. by @csukuangfj in #511
• Fix Byte BPE string results for Python. by @csukuangfj in #512
• Fix #510 by @csukuangfj in #513
• Use high_freq -400 in computing fbank features. by @csukuangfj in #515
• Use NDK 22.1 for android build by @csukuangfj in #518
• Add runtime support for wespeaker models by @csukuangfj in #516
• Support exporting models to onnx from 3D-Speaker by @csukuangfj in #522
• Fix publishing nuget packages. by @csukuangfj in #525
• Add C++ runtime for models from 3d-speaker by @csukuangfj in #523
• Export speaker verification models from NeMo to ONNX by @csukuangfj in #526
• Add C++ runtime for speaker verification models from NeMo by @csukuangfj in #527
• Android TTS APKs for Persian by @csukuangfj in #529
• Fix setting speaker ID for Android TTS Engine. by @csukuangfj in #530
• Add a Persian and a Slovenian model from Piper for Android TTS. by @csukuangfj in #531
• Add Python API examples for speaker recognition with VAD and ASR. by @csukuangfj in #532
• Refactor the UI of Android TTS engine by @csukuangfj in #533
• decoder for open vocabulary keyword spotting by @pkufool in #505
• Change model url from modelscope to github by @pkufool in #538
• Add Android demo for speaker recognition by @csukuangfj in #536

Full Changelog: v1.9.7...v1.9.8

kws-models

Refactor the UI of Android TTS engine (#533)

Release v1.9.7

What's Changed

• Replace Android system TTS engine by @csukuangfj in #508
• Build text-to-speech engine APKs by @csukuangfj in #509

Full Changelog: v1.9.5...v1.9.7

Release v1.9.5

What's Changed

• Fix building wheels for Linux. by @csukuangfj in #484
• Fix CI by @csukuangfj in #485
• Print informative error messages for sherpa-onnx-alsa on errors. by @csukuangfj in #486
• Keep multiple threads from calling into espeak-ng at the same time by @csukuangfj in #489
• Fix whisper test script for the latest onnxruntime. by @csukuangfj in #494
• Release Python GIL in C++ class constructor by @csukuangfj in #493
• Support streaming zipformer CTC by @csukuangfj in #496

Full Changelog: v1.9.4...v1.9.5

Release v1.9.4

What's Changed

• Give an informative log for whisper on exceptions. by @csukuangfj in #473
• convert wespeaker models to sherpa-onnx by @csukuangfj in #475
• Fix releasing go packages by @csukuangfj in #476
• Support playing as it is generating for Android by @csukuangfj in #477
• Fix android tts audio buffer size and fix CI. by @csukuangfj in #478
• Add two GLaDOS TTS models by @csukuangfj in #481
• Play generated audio using alsa for TTS by @csukuangfj in #482

Full Changelog: v1.9.1...v1.9.4

Release v1.9.1

What's Changed

• Remove the 30-second constraint from whisper. by @csukuangfj in #471
• Support distil-small.en whisper by @csukuangfj in #472

Full Changelog: v1.9.0...v1.9.1

Speaker recognition models

This release contains speaker recognition models for sherpa-onnx.

Each model has its own license. Please see the corresponding repository for the specific license of a given model.

Release v1.9.0

What's Changed

• Build building for iOS by @csukuangfj in #430
• Judge before UseCachedDecoderOut by @HieDean in #431
• Build MFC examples for Windows x86 (Win32) by @csukuangfj in #434
• Replace Clone() with View() by @HieDean in #432
• Refactor CI scripts about building wheels by @csukuangfj in #436
• support nodejs by @csukuangfj in #438
• Add Swift API for TTS by @csukuangfj in #439
• Text-to-speech for iOS by @csukuangfj in #443
• Lock before push_back the deque for thread safety by @HieDean in #445
• Update to onnxruntime 1.16.3 by @csukuangfj in #446
• Fix reading tokens.txt on Windows by @csukuangfj in #448
• Fix nodejs on Windows by @csukuangfj in #450
• Release GIL to support multithreading in Python websocket servers. by @csukuangfj in #451
• Support piper-phonemize by @csukuangfj in #452
• Use piper-phonemize to convert text to token IDs by @csukuangfj in #453
• Fix CI by @csukuangfj in #456
• Play generated audio as it is generating. by @csukuangfj in #457
• Break text into sentences for tts. by @csukuangfj in #460
• Support playing generated audio as it is generating for MFC. by @csukuangfj in #462
• Fix building for .Net by @csukuangfj in #463
• Use espeak-ng for coqui-ai/TTS VITS English models. by @csukuangfj in #466
• Support Ukrainian VITS models from coqui-ai/TTS by @csukuangfj in #469
• Release v1.9.0 by @csukuangfj in #470

New Contributors

• @HieDean made their first contribution in #431

Full Changelog: v1.8.10...v1.9.0