Hello and welcome to the Monash DeepNeuron team training! At MDN we regularly work on a broad range of projects, and so it's extremely important for everyone to have a basic understanding of how deep learning works, and how to apply it. All you'll need is a basic understanding of PyTorch and we'll teach you the rest 😀.
Although neural networks involve a lot of heavy-duty theory, we're hyper-conscious of how difficult and overwhelming deep learning can feel for a first-time noob (don't worry, we've all been there). Hence, this time around we're heavily prioritising interactive and beginner-friendly content (no rocket science here 😅).
If you're interested in deep learning and want to know how it works/how to use (like a pro) feel free to plunge through our workshops and accompanying exercises!
This year's training is broken down into 5 workshops, each with an accompanying tutorial/exercise. The exercises can either be completed individually or in a small group (highly recommended).
The first 3 workshops are dedicated to basic concepts, with 1 advanced and 1 final applied session afterwards:
- Introduction to Deep Learning with FastAI - slides - tutorial
- Introduction to PyTorch and PyTorch Lightning - slides - tutorial
- Custom PyTorch Models - slides - tutorial
- Advanced Layers - slides - tutorial
- LSTSMs as an Application of Deep Learning - slides - tutorial
Last but definitely not least we have a [complete guide](Project Walkthrough.md) on how to work on your very first project. This will take you through every step right from deciding on what project/data you want to work with to beginning to write your first training loop. Cement your understanding of everything you've learnt about by using it, in one extremely intriguing demo project to show off to the rest of us!
As you go through our workshops there will be certain things we just... don't have time to elaborate on. When this happens, we'll link to external resources which we find useful.
Note that for basic Python skills any tutorial series like this or this should suffice.