Clarity on homework / "what's due"

Could we get a little more clarity on the expectations for what is due for the homework?

I see two relevant looking repos for the fastai user on github. If you could explicitly state which of those is the one we are using for this class in particular (perhaps in the readme.md for each repo), that would ease some confusion:


Also, assuming that the “fastai/fastai” repo is the correct one, I see two notebooks: “lesson1-sgd.ipynb” and “lesson1.ipynb”. Is the task at hand to shift-enter/understand these notebooks? regenerate them from scratch? Get an accuracy score and paste that somewhere? Become conversant in the technolgoy? While those are all good things I could do, what are the expectations for the class.

Finally, do we turn something in to be graded? It sounds like not, but correct me if I’m wrong.

1 Like

You can find homework for Lesson 1 here: Lesson 1 In-Class Discussion

1 Like

Hi Kayla,

No tangible homework assignment for this week. Just get familiar the tools and code from this week’s lesson.

1 Like

How about homework for lesson 2? I’ve gone through the jupyter notebook - is there anything else?

Kaggle competition? (-:

2 Likes

Yup what @yinterian said. E.g. see if you can replicate my notebook from the dog breeds comp - ideally without watching the video at the same time, or only referring to it when you’re really stuck.

2 Likes

It’s been great watching the fast flow of information on the forums during class. However, I would also appreciate explicit clarity around:

  • Notebooks that were covered in class (w/ links!)
  • Suggested readings
  • Homework / suggested work
  • An overview of what will be covered next class
  • Suggested readings for next class

Right now, I feel like the Wikis are more of a paste bin of topics than a clear direction of where to focus our time.

The class has so many fun things to focus on, it’s hard to pick just a few!

BH

We don’t have any suggested readings, at least not yet, (since at least least week we mainly covered stuff that’s only written up in recent research papers) and I’d rather people focus on experiments and projects using the last lesson, rather than looking to the next one.

So the only things really on your list I think we can provide are:

  • Notebooks: lesson1.ipynb (repo details in lesson wiki post)
  • Suggested work: follow this approach with the above notebook, and then do same steps on another dataset of your choice.

We can use #part1v2-beg for more structured and specific learning resources, I think.

6 Likes