nk243547/ASR-Fellowship-Submission

Domaine:

natural language processing

Type de record:

software

Créateur:

nk2

Hôte:

Adapter-based fine-tuning for low-resource Kinyarwanda ASR # ASR Fellowship Challenge: Adapter-Based Fine-Tuning ## Submission Overview Complete implementation of adapter-based fine-tuning for low-resource Kinyarwanda ASR using the provided WebDataset. ## Important Note The dataset (`DigitalUmuganda/ASR_Fellowship_Challenge_Dataset`) contains **only audio data without ground truth transcriptions**. Therefore, **WER cannot be calculated**. ## Quick Start ### Prerequisites - Python 3.8 or higher - 8GB+ RAM recommended - Internet connection for dataset download ### Automated Pipeline (Recommended) ```bash # 1. Clone the repository git clone github.com cd ASR-Fellowship-Submission # 2. Create and activate virtual environment (recommended) python3 -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate # 3. Install dependencies pip install -r requirements.txt # 4. Run complete pipeline (download → train → evaluate) ./run.sh

nmeedg/ASR-Fellowship-Challenge

Améliorez les performances du système de base sur l'ensemble de données Afrivoice_Kinyarwanda healt

nk243547/generative-ai

Building Generative AI Applications course jointly offered by The Open University of Kenya (OUK) and

Savycode10/ArewaDS-Fellowship

This is a repsository for my assignments an learning materials as an Arewa Data science fellow # Ar

abusarke/Ahyel-fellowship-portfolio

An optimized, high-performance web portfolio featuring localized innovation prototypes Hausa Radio P

MYGBM/fatima-fellowship-blindspots

This dataset provides a targeted, interpretable checklist of reasoning failures and blindspots disco

becaola/Hackathon-submission

This is a project which I submitted in a hackathon I participated in, organized by Pollicy Uganda. T

nk243547/ASR-Fellowship-Submission

Domaine:

Type de record:

Visit

Tasks