This project was done with the completion of the fast.ai course. This neural network trained on bing search image data for the four different seasons and was implemented in pytorch and google cloud services jupyter notebook.
The full application is runnable by clicking the binder button.
Its important to understand that the algorithm is only as good as the data used in the learning process and the search criteria 'fall seasons' and etc are fairly subjective depending on what bings search engine api produces. That said, its impressive what it can do especially for fall and winter seasons where color is a big indicator.