Thanks But honestly, I was just trying not to steal the fun of finding the algorithm to use from participants
He also told me that if the data is unpaired, then pix2pix cannot be used, and CycleGAN and MUNIT are recommended.
Someone recommended me to try out Generative Adversarial Networks, CycleGAN, MUNIT, FUNIT, pix2pix and GauGAN for instance. Deeplearning.ai has a course on GANs on Coursera, (which, I believe can be audited free,) with example notebooks covering the above algorithms. I think it would be better to share with others who are working on this problem.
(post withdrawn by author, will be automatically deleted in 24 hours unless flagged)
There are 2 repeated image names in train.csv and 1 repeated image name in submission.csv
Hi! It seems like data downloaded from the AI Crowd cli still have the labelling errors?
Combining each epoch’s result is really helpful
Looking forward to using the amended version on WILDS!
Solution for submission 152464 A detailed solution for submission 152464 submitted for challenge Iceberg Detection
Solution for submission 152419 A detailed solution for submission 152419 submitted for challenge Clouds Removal
Solution for submission 152146 A detailed solution for submission 152146 submitted for challenge Starship Detection
Solution for submission 150882 A detailed solution for submission 150882 submitted for challenge Docking ISS