Workshops, Fires, New computer, and a Datathon!
Fires
It’s been a terrible start of the year for Australia. I’ve been following the news over on The Guardian.
As a result, we’ve decided to donate 50% of our profits (i.e. just less the venue costs and teaching time), from our two workshops in Feb and March. Hopefully this will be a few thousand dollars.
If you want to do more, my friend Milly noted in a nice thread you can also remove your money from banks and superannuation firms that fund fossil fuel. Bank Australia has been my choice. Finally, Gala and I have also gone through our house and found a bunch of things to give away.
Here’s some good links if you’re looking for places to donate:
- GiveNow - How Can I Donate To Help With The Bushfies?
- The Guardian - Do your homework, give cash and avoid scams: How to donate to Australia’s bushfire crisis
Workshops
We’ve got one workshop in Feb, and a second in March! We’re running the 6 Week Technical Deep Learning Workshop for the second time. It starts on the 6th of Feb. And we’re (in the middle of that workshop!) doing the leadership workshop for the 3rd time, on the 16th of March. Details below.
6 Week Workshop
Starts on the 6th of Feb
- TensorFlow 2.0
- Hands-on
- Collaborative
- Autoencoders
- Convolutional neural networks
AI For Leadership
Runs on the 16th of March
- State-of-the-Art
- Risks
- Tips and ticks
- AI adoption framework
New Computer!
We have decided to name it after someone that inspires us to be better. Mona Eltahawy greets you when you log into our new GPU training machine. If you do not know Mona, you should. She is a feminist giant and author. We were lucky to see her speak at Broadside 2019 in December of last year.
The computer is this one from ScorpTec. It cost $4,900. A solid expense, but it’s already proven very useful for doing ML experiments.
The main thing I wanted was space for GPUs. I picked this computer because it had a good starting setup; a pretty powerful GPU (11GB RAM) + lots of fans. I learned that it’s good to pick a GPU with more cooling, as it means it can run faster. Fine. The next biggest card I found, the 24GB Titan was $4300 alone. While it is better to have more GB in one card, I don’t have the requirement for this right at the minute.
I chose a local computer over a cloud one for “ease of experimentation”. Namely, while I love Google Colab for the workshop; it’s hard to run experiments over several days. It also means I can horde datasets; which is great! And, furthermore, it let’s me keep all my training/experiment checkpoints, which is particularly useful for artistic experimentation, but also useful for training provenance/etc.
In any case, if you’re looking for a GPU machine, ScorpTec is a great place to go! Because I bought off-the-shelf, I ordered the machine at 7.30am and it was ready to pick up at 9.30am that same day! So cool.
Datathon for Social Good
Our Community is hosting a datathon in Feb-March.
The aim is to work with a community organisation (Community Hubs, in this case), and get a bunch of people together and work on a problem, based on data. The event is seeking all kinds of different people, from attendees to mentors. There will be technical work to do, but also work around general problem solving and framing, as well as UI and design work!
Key dates:
- 22-Feb - Launch Event
- 23-Feb - “Hack day”
- 20-Mar - Submission deadline
- 26-Mar - Awards presentation
Check out the website for more details: Datathon For Social Good. We’re very excited to be a part of the event!
