- Python 3.7.4, NPM
- Inside of the API folder,
api.py
contains the frontend code written in Flask - The peddie.py file has all of the model inference code, where we pass a string and an image to make predictions on
- All model training is done in our Kaggle notebook, which can be accessed by clicking the link.
- Download the models from our Kaggle notebook (This also has all of our training code for the models)
- Clone our GitHub repo, and put the models inside of the folder
- Execute in terminal
pip install -r requirements.txt
- Finally,
execute npm run flask-start-api
Trained using an Nvidia Tesla P100 provided by Kaggle.