Lesson 2

From Deep Learning Course Wiki
Revision as of 07:47, 10 September 2017 by Apurva (talk | contribs) (Added link to relevant Podcast Episode (4, 5))
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Full notes for this lesson can be found at Lesson 2 Notes

Ask questions and contribute to the Lesson 2 forum thread. Please help by contributing links, help, and answers to common questions to this page.

Here is the lesson video and the Lesson 2 Video Timeline. All the notebooks from the lesson are available in our github repo. Please read our github wiki page if you are new to github or need help accessing the notebooks. The notebooks are:

  • convolution-intro.ipynb - The convolution tutorial notebook used in the introductory lesson presented during the Data Institute launch
  • lesson2.ipynb - the main notebook for lesson 2
  • dogs_cats_redux.ipynb - how to enter the Dogs vs Cats Redux competition, and how to visualize your models correct and incorrect predictions
  • sgd-intro.ipynb - the simple SGD tutorial

The spreadsheets shown in the class are available in the github repo.

This week's assignment

  • Complete the readings below - especially the Stanford CNN pages
  • Complete a submission to the Dogs vs Cats Redux competition if you haven't already
  • Try to get in the top 50% of this competition
  • Try to set up the data, train a model, and submit your predictions for this competition using only code that you write from scratch. If you look at the code provided for the class to get help, don't copy and paste it - but instead read the source code for vgg16.py and utils.py, and the keras documentation, and try to write and debug your own version
  • Enter another competition, selecting from Image Datasets, or create a model using your own data, using the techniques learnt so far

This week's readings

Additional Resources

  • This week in Episode 004 of the Startup Data Science podcast, Apurva, Alex, and Edderic examine the need for creating test and valid folders. In Episode 005, we have a casual discussion about learning rates and derivates, and what excites us most while learning Lesson 2.