Pytorch NLP Recipes

Introduction

Pytorch NLP Recipes is a simple implementation of a deep learning project for NLP, in particular PyTorch. This project contains two problem statements:

• Sentiment Analysis
• Named Entity Recognition

The workflow contains a couple of steps:

1. Ingestion to a datamart (Schema used for training & testing)
2. Building a model
3. Evaluation of a model
4. Deployment (used for inference)

Setup environment

Initialise your Python environment. I use pyenv / virtualenv, and 3.11.1

make setup

Ingest datasets

make ingest

Train

make train

Eval

make eval

Remove environment

make remove