language detection from text for English and Luganda

This will include text used and the notebook

PS:

The large vocab csv file needs alot of RAM to compute , therefore I propose using the smaller vocab csv

The notebook identified here was run using google colab

You can find the csv file for either the larger or smaller vocab in the data folder

To run this

Install the requirements in a virtual environment using

Then you can start the app using

 gunicorn app:app

You can access the endpoint using

http://127.0.0.1:8000/predict?text="Obumenyi bw’amateeka bufuukidde Pulezidenti n’eggwanga ekizibu."

Origin of data used

A parallel corpus dataset

Origin of notebook used

notebook on kaggle