Using Deep Learning For Image and Sequence Analysis (ECCB2020 tutorial)

Shorten Link: http://bit.ly/eccb2020

Slides: https://docs.google.com/presentation/d/13S9ljSIQglEPihzcj7VoVl8ptsP9MZ_wEJ6QUnkErSk/edit?usp=sharing

Click on the links below to open examples in Google Colaboratory

• 01_one_neuron.ipynb - Simple regression problem

• 02_MNIST.ipynb - Digit recognition

• 03_ImageNette.ipynb - Training network on a small ImageNet subset

• 04_Transfer_Learning.ipynb - Now, it is your turn to find interesting image recognition problem and train the network

• 05_Sequence_Generation.ipynb - Generate text or genomic sequence

• 06_Sequence_Classification.ipynb - Classify texts or genomic sequences

Links to fast.ai resources:

While we are in no way affiliated with fast.ai, this tutorial would never come to life without

• Practical Deep Learning for Coders course
• Deep Learning for Coders with fastai and PyTorch: AI Applications Without a PhD book, also you can peek into its draft on GitHub
• Fastai 2.0 library - also, see blog posts explaining the idea of its layered API architecture and recent developements

Acknowledgements

We are indebted to the helpers that reviewed the content and were answering questions on the forum: Vlastimil Martinek, David Cechak, Karla Fejfarova and Ondrej Vaculik. Most of our materials build on fastai library and awesone Practical Deep Learning for Coders (notebooks 3-4 are practically taken from the course). The second notebook is highly inspired by keras MNIST example from Tensorflow Dev Summit 2018. We would like to thank fastai and pytorch creators for giving us sucha powerful tool and Google Colaboratory for GPU runtime (running DL tutorial for 70 participants and having no technical glitch is some kind of a miracle).