Sven van der Burg

Our current approach of balancing the data during data generation is picking an inchikey, than looking for a pair in a randomly sampled similarity bin, and iteratively widening the bin...

I think we should maybe justify the rather complicated way we do training data generation. There is 2 things we should show I think: 1) that balancing improves the test...

I saw at least 1 case where this is not the case, namely `same_prob_bins` by default is [(0, 0.5), (0.5, 1)], whereas in the paper we use more and smaller...

outlook [refers to cifar-10](https://github.com/carpentries-incubator/deep-learning-intro/blob/main/episodes/6-outlook.Rmd#L97) in a way that I think should be swapped to dollarstreet

the change to the dollar street data also opens a really great opportunity to do bias evaluation and talk about the importance of having many dimensions of diversity in a...

[the index](https://github.com/carpentries-incubator/deep-learning-intro/blob/main/index.md?plain=1#L40) refers to the content as a workshop, but I think "lesson" is the more correct Carpentries term. the [later use of "workshop"](https://github.com/carpentries-incubator/deep-learning-intro/blob/main/index.md?plain=1#L44) I think is correct.

Add more terms to glossary

1

We can add a few more terms to the glossary, because we have to submit a PR to glosario I am keeping a list here that we can pick up...

Add section on hyperparameter tuning

2

Fixes #351 @reviewer: This is now in episode 4, I was thinking to maybe put it together with transfer learning in the new episode 5 (See #349). Episode 5 could...

Now that the lesson is stable we can remove the call(s) for beta pilot testers.

Look at https://github.com/carpentries-lab/reviews/issues/25, see if there are any minor comments that we missed and can still improve.