Self-Supervised Learning Challenge 2019


Fail to submit SVM-Lowshot file (4)
Unable to reproduce benchmark results mentioned in the paper (5)
Can we use different self-supervision settings in different tracks? (6)
Are there any limitations on the backbone architecture? Is it limited to AlexNet and ResNet-50, or we could use any architecture with any capacity? (1)
Are there any limitations on the backbone architecture? Is it limited to AlexNet and ResNet-50, or we could use any architecture with any capacity? (1)
Is it okay to use videos dataset and optical flow computed in raw videos using traditional (e.g., Lucas-Kanade) or advanced (flownet2) methods? (1)
Is it okay to use Raw Videos dataset? (1)