Mercari Item Classification

This repo contains a dummy project created during Build@Mercari 2021 to learn about the Deployment of ML Models

Steps to run the application locally after cloning the repo:

Make sure you are inside the root directory mercari-item-classification
Download dataset from https://www.kaggle.com/c/mercari-price-suggestion-challenge
Put the train.tsv and test.csv under mercari_model/datasets/
Run python -m venv env to create virtual environment, then activate it source env/Scripts/activate
Run pip install -r requirements.txt to install required packages
Run tox -e train to train and save the pipeline for later inference
Run pytest or tox to run pytest scripts make sure all APIs are working
Run export FLASK_APP=flask_apis to set which module contains the Flask API
Run flask run to serve the api, or run python wsgi.py
Test using any frontend app or Postman to try the APIs

Next todo:

Provide feedback