ResNet
Era: The Revival · 1990s–2010s
Kaiming He et al. proposed ResNet, solving the training difficulty of deep networks through skip connections. ResNet enabled training hundreds of layers.
Kaiming He
Sources and support
Deep Residual Learning for Image Recognition
Microsoft Research / arXiv · Source type: paper
He and colleagues introduce residual learning and report ImageNet results.
Published: ·Accessed: