nlp-handson

This repo contains starter code for NLP-related job interviews. There are two parts that rely upon one another:

1. a training script (train.py) that downloads a dataset from the HuggingFace Hub, trains a simple bag-of-words classifier and saves both the classifier and the feature extractor
2. a streamlit app that creates a simple form and uses the classifier and feature extractor saved previously to classify answers into one of three classes

Setup

1. Clone your fork.
2. Not strictly relevant, but highly recommended: create a virtual environment for project dependencies, e.g. python -m venv venv
3. Install requirements: pip install -r requirements.txt
4. Run training script: python train.py
5. Run streamlit app: streamlit run app.py

Once the above steps are done and things are working, you are ready for the interview :-)