What? ML Sessions.
When? 9th, 10th and 12th April 2018.
Where? CRC 102.
Greetings from the Computer Vision and Intelligence Group!
We are organising a three day Machine Learning session on the 9th, 10th and 12th of April. The sessions aim to provide a fairly comfortable start to ML, while getting you upto speed with covnets and SVMs.
We’ll be delving into both the libraries popularly used to deploy architectures, as well as a bit of math behind it. By the end of the sessions, you should be able to train your own neural net and understand (to some extents
Laptops will be needed for the sessions (no kidding), and the software and the other pre-requisites are enlisted below.
- Session One: Collab [link] (https://colab.research.google.com/drive/1lTye1Dv2VJ3vqstdfOElD1FcWjTmTLGM)
- Session Two: Collab [link] (https://colab.research.google.com/drive/1SyLP06KH1K-AQzbnkUXhQKnt5hESSBsk)
- Session Two: Collab [link] (https://colab.research.google.com/drive/1b4AAqd1AYII8vK3YUjJMEHc8XCxNnW-U)
Note: We will be shortly updating a script file for installs. This will however make use of pip and will work for Linux and macOS users only.
Operating System: Preferably Linux or macOS(formerly OS X). You could weather the battle with windows however, consider that WSL for windows and conda have got better over the years. Regarding linux, please stick to LTS releases (hello Ubuntu), since tensorflow can be troublesome on the shorter releases.
OpenCV: OpenCV 3.3+
Python: Python 3.5. Python 2.7 is fine too, but you’ll have to watch out for subtilities during the sessions.
Tensorflow: 1.4+ (CPU or GPU).
Numpy and Scipy.
Lastly, we would need you to have at least pip for package installs.
In case you are installing ubuntu (or any other distro) for the first time, you might find it convinient to head over here for a comprehensive setup script. Due credits: Nikhil Krishna
We’ve been putting together a fairly comprehensive post for getting started with deep vision, RL and NLP. You can find it here.
Do reach out to us if you need any support. You can post any issues you may encounter on the group forums.
Join the WhatsApp group to stay in touch: link [https://chat.whatsapp.com/EEN2EpQbJCk8wy2kaKR8ZQ] .