My First Deep Learning Notebook

I recently started fastai’s Practical Deep Learning for Coders course where in the first two lectures you are encouraged to build and deploy a deep learning application.

Deploying a deep learning app before learning in-depth about neural networks and deep learning follows fastai’s co-founder Jeremy Howard’s philosophy that in order to learn a subject you must tackle problems that are actually relevant to the content you are learning.

The deep learning app which is hosted on Binder and can be found by clicking the badge below, was trained to classify pictures of curling houses and actual houses. The model was trained in the Kaggle notebook, here. The Github repo which is used to build the Binder server can be found here.