sherpa-onnx
sherpa-onnx copied to clipboard
k2-fsa
Speech-to-text, text-to-speech, speaker diarization, speech enhancement, source separation, and VAD using next-gen Kaldi with onnxruntime without Internet connection. Support embedded systems, Android...
使用 sherpa-onnx + SenceVoice 的onnx vad-with-non-streaming-asr.py 就是 使用 microphone 去 asr 现在是这样 原来 SenseVoice 是有情感输出的。 ban_emo_unk：禁用 emo_unk 标签，禁用后所有的句子都会被赋与情感标签。默认 False 是否可以像 SenseVoice 那样，能有开关，输出 情感标签？ 毕竟， 情感输出也是 Sensevoice 的一大特色。谢谢。
When running whisper **medium** (int8 / normal) on DirectML provider it failed. on CPU it works. same build. *The error* ```console Non-zero status code returned while running MemcpyToHost node ```...
Hi there! I'm trying to embed sherpa in an android application using c# xamarin. org.k2fsa.sherpa.onnx v1.10.17 package works well on windows and linux. On Android there is a problem with...
测试中文的唤醒词 git clone https://github.com/k2-fsa/sherpa-onnx cd sherpa-onnx mkdir build cd build cmake -DCMAKE_BUILD_TYPE=Debug .. make -j6 运行测试命令 ``` ./sherpa-onnx-keyword-spotter --tokens=/mnt/data/speech/keyword-spot/models/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01/tokens.txt --model-type=zipformer --encoder=/mnt/data/speech/keyword-spot/models/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01/encoder-epoch-12-avg-2-chunk-16-left-64.onnx --decoder=/mnt/data/speech/keyword-spot/models/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01/decoder-epoch-12-avg-2-chunk-16-left-64.onnx --joiner=/mnt/data/speech/keyword-spot/models/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01/joiner-epoch-12-avg-2-chunk-16-left-64.onnx --provider=cpu --num-threads=2 --keywords-file=/mnt/data/speech/keyword-spot/models/sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01/keywords.txt /mnt/data/speech/output_1.wav ``` 错误信息： KeywordSpotterConfig(feat_config=FeatureExtractorConfig(sampling_rate=16000,...
Shallow fusion can be too slow for online cpu inference. Added an option to use classical LM rescore instead. - Rescore implementation based on https://github.com/k2-fsa/sherpa-onnx/pull/133 - Shallow fusion enabled by...
https://github.com/k2-fsa/sherpa-onnx/pull/1224 added support for online punctuation models. However, only C++ is supported. Help from the community is appreciated to support the following languages: - [ ] 1. C - [x]...
Heya, First off, many thanks for working on bringing this package to Dart. I have been working with this package since several months ago in Flutter in my own way...
max-active-paths的值设置为大于1时（如2、4、10）说一遍关键词，关键词会被识别到两次，值越大出现的概率越高； 例如在控制台输入如下指令： python3 ./keyword-spotter-from-microphone.py \ --encoder /Users/gree/Desktop/sherpaonnx20240723/kws-zipformer-gigaspeech/encoder-epoch-12-avg-2-chunk-16-left-64.onnx \ --decoder /Users/gree/Desktop/sherpaonnx20240723/kws-zipformer-gigaspeech/decoder-epoch-12-avg-2-chunk-16-left-64.onnx \ --joiner /Users/gree/Desktop/sherpaonnx20240723/kws-zipformer-gigaspeech/joiner-epoch-12-avg-2-chunk-16-left-64.onnx \ --tokens /Users/gree/Desktop/sherpaonnx20240723/kws-zipformer-gigaspeech/tokens.txt \ --max-active-paths 10 \ --num-trailing-blanks 0 \ --keywords-file /Users/gree/Desktop/sherpaonnx20240723/kws-zipformer-gigaspeech/keywords.txt \ --keywords-score 1.5 \ --keywords-threshold...
