Hi the organisation team,
I am trying to submit my prediction results to the server, but I found some submission rules for the prediction challenge:
- Unlike previous challenges, the leaderboard will be ranked according to performance on the nuScenes val set. This is because we cannot release the annotations on the test set, so users would not be able to run their models on the test set and then submit their predictions to the server. To prevent overfitting on the val set, the top 5 submissions on the leaderboard will be asked to send us their code and we will run their model on the test set. The winners will be chosen based on their performance on the test set, not the val set.
- Every submission to the challenge must be accompanied by a brief technical report (no more than 1-2 pages) describing the method in sufficient detail to allow for independent verification.
Are there rules still valid after the the ICRA 2020 workshop? What shoud I do if I want to submit my prediction results to the leader board to have a benchmark after that?