This is a collection of examples of using deep learning to solve different types of problems. I am using it as a way of getting practical experience while I follow along with the (fantastic) http://course.fast.ai/ course and the deeplearning.ai Deep Learning specialisation https://www.coursera.org/specializations/deep-learning. I will update the examples with any new methods I discover of doing each type of task, so that hopefully the results will improve over time.
The /experiments
folder contains things that may be interesting, but are not necessarily explained or guaranteed to run!
I'll use publicly available datasets so all the examples are reproducible.
Here are all the examples and associated test results.
Multi-label classification:
- Seedlings dataset: 97.64% accuracy on the test set