Hello from DeepMind! We've been watching from a distance and have been very happy to see the constant growth of leaderboard submissions. DeepMind is a great believer in leaderboards as a way to make progress on hard problems - whether that's measuring the score in an Atari game, the error in folding a protein or, as in this (possibly equally hard!) case, the accuracy of a machine learning model on a test set. We hope that the tutorial we provided was useful in getting an initial model running in Jax and would love any and all feedback:

* Was the tutorial easy to run on a colab instance?

* Did it cover enough to let you hit the ground running?

* Was the code easy to follow?

* Was anything particularly confusing on a first read?

DeepMind and Zindi are planning a webinar to be shared soon. We look forward to answering any questions you may have for us or going through the starter notebook.