Jen eta provi fari parolrekonilon uzante la Vosk parolrekona modelo.01 (vosk-model-small-eo-0.22)
Vi povas elŝuti de https://javabog.dk/privat/EsperantoParolRekono.apk
Note: This application is just a proof of concept for now
LocalSTT is an Android application that provides automatic speech recognition services without needing internet connection as all processing is done locally on your phone.
This is possible thanks to:
- a RecognitionService wrapping the Vosk library
- a RecognitionService wrapping Mozilla's DeepSpeech library
- an Activity that handles RECOGNIZE_SPEECH intents amongst others
The code is currently just a PoC strongly based on:
LocalSTT should work with all keyboards and applications implementing speech recognition through the RECOGNIZE_SPEECH intent or Android's SpeechRecognizer class. It has been successfully tested using the following applications on Android 9:
You can download a pre-built binary with Vosk models for:
and also with DeepSpeech models here:
- Catalan: here.
If you want to use the application with your language just replace the models in app/src/main/assets/sync/vosk-model/ with a package from https://alphacephei.com/vosk/models and rebuild the application.
- git clone https://github.com/ewheelerinc/LocalSTT.git
- ./gradlew build
- ./repack-n-sign.sh ./app/build/outputs/apk/release/app-release-unsigned.apk
- You might need to update paths and keys in this script for your use.
- Does not work with Google's keyboard "GBoard".
- Not all record applications read the voice text properly, there must be another way---and if you know how, it is probably a trivial fix.
- DeepSpeech models were removed, they didn't build! Maybe it can be fixed?
- Support query alphacephei.com and suppport selection+download of optional models. Then this apk can be packaged without a language (much smaller!).
Nota: Aquesta aplicació de moment només és una prova de concepte
LocalSTT és una aplicació per Android que proporciona reconeixement automàtic de la parla sense necessitat de conexió a internet ja que tot el processament és local al mòbil.
Això és possible gràcies a:
- un RecognitionService que utilitza la llibreria de Vosk
- un RecognitionService que utilitza la lliberia de Mozilla Deepspeech
- una Activity que gestiona intents RECOGNIZE_SPEECH entre altres
El codi és actualment una prova de concepte i es basa fortament en els següents projectes:
LocalSTT hauria de funcionar amb la majoria de teclats i aplicacions que implementen la funció de reconeixement de veu a través d'un intent RECOGNIZE_SPEECH o directament fent servir la classe SpeechRecognizer d'Android. Ha estat provada amb èxit fent servir les següent aplicacions en un terminal Android 9:
Us podeu descarregar un APK que inclou models de Vosk i DeepSpeech pel català aquí.