How will the second round differ from the first?
With each round the dataset size will increase. The motivation is to beat the last achieved score. You may be able to use the same model from round 1 in round 2 if the number of classes remain the same but with slightly less mAP. If the number of classes differs then you’ll have to retrain the model.
2 Likes
At this point, the number of classes are scheduled to remain the same. The number of training images and associated annotations will increase nevertheless.
2 Likes