FAST Offline Speech Recognition with One of the fastest Whisper implementation that works flawlessly real-time cross platform with JAVA. It works on Linux , Windows , OSX , Android even on IOS ( LibGDX ROBOVM near future ) Its model is far more optimized.
The quality may be increased with a cost of losing performance. We will update for that.
The library works on Java 8 , 9 , 11 , 19 , 21 , 24+
This is the best Java Whisper Library on the internet. Dont lose time for Whisper JNI implementations Dont try to search more and its completely FREE
It can supports CUDA,FAST CPU,AMD ROCM,INTEL Its ram consumption for X86 is just 1.5GB MAX on JVM !!!
We tried to minimize garbage collection usage as far as we can. We will upgrade this library more as soon as we have time more.
We will try to support library as possible to make more usable for Java.
TESTING :
For Windows UTF-8 CMD:
chcp 65001
WINDOWS AND OTHERS:
git clone https://github.com/eix128/WhisperJET
cd WhisperJET
mvn package
cd target
mkdir models
copy ..\src\main\resources\models*.* models</code>
mkdir testWavs
copy ..\src\main\resources\testWavs*.* testWavs</code>
java -jar WhisperJET-1.0-jar-with-dependencies.jar
USE OF LIBRARY :
git clone https://github.com/eix128/WhisperJET
cd WhisperJET
mvn install
copy models folders from : src\main\resources to your project
You can checkout for codes :
Simple wav for 30 second test
WhisperJET recognizerJava = new WhisperJET();
byte[] bytes = Main.readWavNClipMax30("testwav.wav");
long preCheckTime = System.currentTimeMillis();
String translatedText = recognizerJava.recognize(bytes , languageCode );
We will add realtime examples and support also.
Stay tuned!
Reference Works :
https://github.com/niedev/RTranslator
Feel free to ask any questions :
[email protected]
Donate By BTC Address :
1KuREYKb1fnnVJEBMXd1P7XmkjZcrTY3A3
REFERENCE WORKS:
https://github.com/niedev/RTranslator