Could someone share the link if I missed it. Thanks!

Congrats to all winners! Interesting and impressive approaches!My approach utilized 100 classes from train data (without the extra train data). Final model is... See more

We really enjoyed the competition and particularly spent a lot of time researching how state of the art facial recognition systems are built. We learnt a lot an... See more

I made visualisation of places where most activuty happens during turtle analysis by neural net. It was made for 2 different nets. One is DenseNet121, other is... See more

Congrats to the winners, and thanks to Deep mind and @Zindi for hosting. I had a great time in this competition, I however noticed many folks were unable to cro... See more

Hello Zindi,I'm facing a serious issue where I can't pick my best two submissions as one of my previously selected submission is not appearing in the submission... See more

Hello again from DeepMind! Only 3 days are left before the competition ends, which means there is still time for final improvements! We hope you have enjoyed th... See more

I was wondering how I can correctly validate my models and it makes sense to validate using images of the most recent turtle captures as the validation set (hav... See more

At inference time, we need to handle the case when there is a turtle image fro, unseen class (which is not in the 100 classes on which we train). How do we hand... See more

We are given Extra images data and the size is a little big. Therefore, we should use this data for better score. Then, I have a question. How can we use new t... See more

Hi! I was thinking that for those interested. It would be nice to share ideas on what has been working and what does not. Now that the validation seems closer t... See more

Hi all,There were a lot of disscusions about low public scores despite high validation, and there's my two cents about this phenomenon.I assume that some mess h... See more

Hello Zindians,I joined this challenge early this week and I started creating a baseline where my local validation wasn't that bad compared to what I've got in... See more

We know that the CV is so different from the public LB. I am wondering if @Zindi could hint us how the private LB looks like? The top scores would be still arou... See more

Most of you have expressed concerns about the cross-validation scores being way better than the leaderboard scores; thank you for raising the concerns.Most of t... See more

I'm sorry for the inconvenience but in this competition there is something that is not right and maybe together we can solve it.I have spent time creating a goo... See more

EDIT: I didn't set things up for perfect reproducibility and those who have re-run have seen lower scores in practice. So take the 0.064 with a pinch of salt un... See more

I get that the temptation is to treat 'turtle_id' as a label and approach this as a classification problem but that feels a little weird given the context. Addi... See more

Are we allowed to use image location during the training phase?For rules, we can't use metadata but in the test set the location is also provided. Could someone... See more

I notice that there is a great gap between CV & LB. I have seen some great discussions about that topic. So this gap can be explained by the fact that less... See more

There are 100 unique IDs in train csv while extra + train has close to 2.2k unique IDs. I don't understand why the extra csv isn't just added in as train csv. I... See more

Join us on Thurs Feb 10 17:30 CAT for a live webinar with researchers from DeepMind. They will walk through the starter notebook for the #TurtleRecall challenge... See more

My submission as well as sample_submission end up with the same message. I have all this image_ids in my csv. What's wrong?

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 bel... See more

Hi there, are we allowed to assume the image location is available at test time or is this part of the meta data that we should not use?