Deep Learning with Keras & Tensorflow: from Zero to Hero in 3 Hours
a workshop with Karla Fejfarová & Petr Šimeček
Deep Learning is everywhere. It beats us in Go and Chess, translates our texts between languages, turns steering wheels of self-driving cars and tags Not-Hotdogs on out photos.
We will introduce the basic concepts and demonstrate them on two examples: dogs vs. cats image classification and text generation from Nietzsche’s writings. Moreover, we provide similar datasets for you to practice and, in the end, deploy the trained models as a simple API service.
Because we will be working in a cloud, all you need is a laptop and a decent web browser (not IE 6.0). Beginners are welcome. Ph.D. degree in machine learning definitely not needed. However, a familiarity with Jupyter Notebooks and some experience with numpy module might be useful (you can easily get it from Udacity/Datacamp/… online materials).
This workshop is suitable for both beginner and advanced Pythonistas.
I'm a researcher at the Institute of Molecular Genetics by day and a Python girl by night (also early mornings, due to fast.ai fellowship and time difference between Prague and San Francisco). Most of my data look like a mouse skeleton. Deep learning is one of my data processing tools.