If object detection is performed on all labels, or if strong data extensions are used, ImageNet pre The -trained model can adversely affect the accuracy. On the other hand, if they use Self-training (Noisy Student) to learn beforehand, found that even in those cases, the accuracy is improved.However, the accuracy is degraded using SimCLR as self-trainig method. This may suggest that just "Noisy Student is great method".
TL;DR
If object detection is performed on all labels, or if strong data extensions are used, ImageNet pre The -trained model can adversely affect the accuracy. On the other hand, if they use Self-training (Noisy Student) to learn beforehand, found that even in those cases, the accuracy is improved.However, the accuracy is degraded using SimCLR as self-trainig method. This may suggest that just "Noisy Student is great method".
Why it matters:
Paper URL
https://arxiv.org/abs/2006.06882
Submission Dates(yyyy/mm/dd)
Authors and institutions
Methods
Results
Comments