BrDhaker/TunisianASR

Tunisian Automatic Speech Recognition Mobile App # TunisianASR TunisianASR is an Automatic Speech Recognition (ASR) mobile app specifically designed to transcribe Tunisian Arabic speech. The Transcription can be made using a real time recording or an audio file from the device local Storage. ## Technologies - **WavLM Model**: Transformer-based deep neural network for speech recognition. - **KenLM**: 4-gram language model for enhanced transcription accuracy. - **Gradio**: Interface for creating API and web-based demos. - **Flutter and Dart**: For developing the mobile app. - **Hugging Face**: Hosting and deployment platform for the ASR model. ## Screenshots ### Dark Mode *App in Dark Mode* ### Light Mode *App in Light Mode*