Dear organizers,

Thank you all for organizing this great challenge. We have observed that our methods don't work well on some cases (maybe they are not similar with training data), so is there any possibility that we get access to the testing set (or part of it) to further enhance our methods?

Best regards, on behalf of team IGIP.