JONAHKYAGABA/African-Language-ASR-Model-Evaluation

African Language ASR Model Evaluation # 🌍 African Language ASR Model Evaluation This project provides a comprehensive framework for evaluating Automatic Speech Recognition (ASR) models on low-resource African languages such as **Amharic**, **Oromo**, and **Shona**. It computes evaluation metrics like **Word Error Rate (WER)** and **Character Error Rate (CER)** and supports batch inference on Hugging Face models. --- ## 🎯 Objectives - Load ASR models from Hugging Face (Whisper, MMS, Wav2Vec2, etc.) - Preprocess African-language datasets - Run automatic speech recognition inference - Compute and log WER and CER metrics - Save detailed CSV reports of predictions ---