Hey Data Explorers,
This is the regular discussion session for week 6. Thanks to the awesome folks at Planswell, we get to build neural networks in their office!
What to expect in this session:
1. Get to know each other
2. Go through notebooks 6 and discuss experiments that worked or didn't work
3. Share tips on how to "fully digest" lecture 6 before next week.
What's required to attend this session:
1. intermediate Python (the course website says 1 year of coding. Basically the more coding experience you have the faster you will learn, but it's recommended that you know all the basics in an intro to Python course, or there will be lots of catching up to do!)
2. high school math
3. a laptop
4. completed watching lecture 6 (or you'll spend time watching it during the session)
5. be friendly :)
General Series Format:
There are at least 2 sessions (Mondays and Thursdays) per week for 7 weeks (Oct 29 - Dec 13). Currently the setup is the following:
Mondays: Watching the lectures together (~ 2hrs)
Thursdays: Discussion lead by Xu Fei and then coding, as well as for those who want to catch up with the Monday lecture.
Xu Fei went to USF earlier this year and completed the Fastai part 2 Cutting Edge Deep Learning for Coders in person. This time he'll share as many things that helped him in the course as possible, and hope to learn from everyone else.
Since this is a pretty intense course, we can also add a few sessions on weekends (location TBD) for those who need additional help, or whoever wants to work on solving problems together. This will be determined by everyone before the end of Thursday.
Why is this course offered for free?
We offer it for free because the creators of the course keep it that way for the online version. Please read more in the "Additional Info" section below for the mission of Fastai, and you will find out more once you join the course ;)
This course is the 3rd run of Practical Deep Learning For Coders, Part 1 offered by Fast.ai. It uses the brand new Fastai v1 library, based on PyTorch 1.0 released at the beginning of October 2018. You can think of Fastai to PyTorch as Keras to TensorFlow, or take a quick look at this quote
If you haven't heard of the Fastai deep learning course, please take a look at their previous courses:
An inspiring TEDxSF talk by Rachel Thomas why she and Jeremy Howard started Fastai:
A very recent article covering Fastai in The Economist.
Currently the course is taking place at University of San Francisco every Monday evening (started on October 22nd for 7 weeks). Although the deadline to register for the live online version has passed, we fortunately got permission from the instructor Jeremy Howard to share the latest course content with our group members offline, so motivated learners from Toronto can access the raw/live version almost in sync (with 1 week of delay) with the San Francisco students. The polished version will be open to the public in January 2019.