Version released 0.3 a library vosk for local recognition of continuous speech, supporting Russian language.
An APK package has been prepared for the Android platform while for Linux you can use thePython library example usage (), with performance sufficient for operation on Raspberry Pi boards.The library operates on an enhanced engine
Kaldi The language model.
takes up only 50MB and works more accurately than DeepSpeech (model size over 1GB). Supported languages: Russian, English, German, French, Chinese. Support for Spanish, Hindi, Arabic, and Portuguese is expected.
Virtual hosting or virtual server — what to choose?
Source: linux.org.ru
