SalamaBsd/Speech-to-Text-and-Text-to-Speech-system-for-Moroccan-Arabic-Darija-using-Deep-Learning

This project implements a voice system for Moroccan Arabic (Darija), combining Whisper for speech-to-text (STT) and SpeechT5 for text-to-speech (TTS). 1. Both models are adapted to a custom Darija Latin dataset. 2. The goal is to enable natural interaction with local dialects using deep learning.