/Fastai-Deep-Learning-From-the-Foundations-TWiML-Study-Group

Review materials for the TWiML Study Group. Contains annotated versions of the original Jupyter noteboooks (look for names like *_jcat.ipynb ), slide decks from weekly Zoom meetups, etc.

Primary LanguageJupyter Notebook

TWiML Fastai "Deep Learning From the Foundations" Study Group

This repo is to share annotated Jupyter notebooks, lecture slide decks and other curriculum materials that I'm creating for the TWiML Fastai "Deep Learning From the Foundations" Study Group aka "Practical Deep Learning for Coders Part 2"

Annotated or otherwise modified notebooks bear the suffix '-jcat.ipynb'.

The study group meets remotely via Zoom on Saturdays at 8:45 AM Pacific time, from July 6, 2019 through November 30, 2019. We'll cover Lessons 8 - 12 of the course, which are associated with the Python / PyTorch Fastai library. We will not cover lessons 13 and 14, which provide an introduction to the new Swift / TensorFlow implementation of the Fastai library. We communicate on the on the #fastai_dl channel of the TWiML Slack Group. Our use of the Slack and Zoom platforms is by courtesy of Sam Charrington, of TWiML (This Week in Machine Learning).

An up-to-date schedule is maintained at https://docs.google.com/spreadsheets/d/139XMd3moRb1C2No1XF5zN11Ivqm1XXXhEWNu2wcfTZc/edit#gid=339739437

Join the TWiML Fastai study group here: https://twimlai.com/twiml-x-fast-ai/

Here is the key to optimizing your Fastai course experience: BE MORE THAN A SPECTATOR! Even if you intently watch and study all the videos, unless you're also learning from the course notebooks by spending time with them, running them and playing with them, you are a spectator, and you will miss out on getting the full benefit of the course.

I've found that the primary barrier to participation in the Fastai courses is the task of setting up infrastructure to run the course notebooks. There are many possible ways to accomplish the task. If you have a computer with a GPU, you can run the notebooks locally. Otherwise you can run the notebooks on Kaggle, on Google Colab, or on other cloud platforms such as paperspace, AWS, and crestle.

Currently the best cloud alternative for most people is Google Colab, which provides free access to CPUs, GPUs and even TPUs (Tensor Processing Units)! I have created a Jupyter notebook to guide you through the process of setting up to run the course notebooks on Google Colab; you can run it directly from this repo via the URL

https://colab.research.google.com/github/jcatanza/Fastai-Deep-Learning-From-the-Foundations-TWiML-Study-Group/blob/master/initialize_fastai_dl2_colab_jcat.ipynb

If you plan to participate in the course, please set up infrastruture to run the notebooks as a first priority!

Hope to see you at the next meetup!