This course was built to answer some questions that I encountered when I started my deep learning journey that I didn't find a good resource for: what is it like building a deep learning project? What are the challenges? How much time does it take? What will take most of my time? What are the available tools? How easy is it to build a performant model?

We answer these questions in this course by implementing a basic image classifier using modern deep learning tools and taking our time in understanding how this basic image classifier can evolve and what challenges we might face when moving into the real life outside of structured courses.

The implemented image classifier is trained on a Kaggle dataset of sports so that it can differentiate between 73 different classes of sports given only an image. The used model is EfficientNet and the deep learning framework used is Pytorch Lightning. The implementation includes implementing the data loader, the model that wraps EfficientNet, and a trainer.

What is primarily taught in your course?

Two things are taught in this course: implementing a basic but complete image classifier program from scratch and getting to know the development flow, challenges, tools, & opportunities involved in developing a deep learning project.

Who are your target students?

This course is targeted towards engineers who don't have a background in deep learning and want to know what it is like building deep learning pipelines.