This repo contains codes about loading audio data, training wav2vec2 model with wolof language dataset
-
wolof contains the wolof alffa train datasets 01, 02, 04, 07, 08, 09
-
customWav2Vec2_notebook is the notebook used to train and push the model to the hub
-
The file requirements.txt contains all packages to install
-
The module utils.py defines the class for the fastAPI inference
-
main.py constructs the fastapi interface
-
gradio_demo and flagged are for the gradio interface
-
key and key.pub are the private ant the public key for docker
-
Dockerfile is to build docker images and run containers
-
captures des results contains all screenshoots of the results obtained for the training and the interfaces (fastapi, gradio, docker-build, docker-run, docker-inspect)