practical-ml-seminar

Supporting code for Practical ML for development research seminar

Setup

python3 -m virtualenv venv
source venv/bin/activate
pip install -r requirements.txt

Additionally, if you wish to make a new export from the IATI datastore API, you must create a .env file with an API key you can get from the IATI developer portal (https://developer.iatistandard.org/).

To make a new export from the IATI datastore API

python3 datastore_api.py

Notebook order

First, run train_classifier.ipynb either locally or on a remote python kernel. Second, you can run inference_classifier.ipynb to test the model you just trained.

Last, if you want to experiment with zero-shot classification, zero_shot.ipynb

Google Colab

A Google Colab implementation of this example is available here for the pre-trained classifier. and here for the zero-shot classifier.