I’ve had heaps of fun with round 1. This was my first time learning how to use git (let alone git-lfs). I learned how to use graphics hardware to my advantage (and how to keep my house warm in the cold nights!). Most of all, I’ve really appreciated the community support y’all have given each other
What does this new round entail? Personally, I’d love to see a cleaner dataset. I believe that the more accurate the human-made segmentations are, the better a model with embeddings could perform (e.g, HTC).
Any news on other Prizes? It could be interesting using distributed computing/ a cloud service to quickly train a model.
I am very excited for the future of this challenge!