ECA-Net: Efficient Channel Attention for Deep Convolutional Neural Networks

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetECA-Net (ResNet-152)#755Top 1 Accuracy: 78.92%Number of params: 57.40MGFLOPs: 10.83
image-classification-on-imagenetECA-Net (ResNet-101)#769Top 1 Accuracy: 78.65%Number of params: 42.49MGFLOPs: 7.35
image-classification-on-imagenetECA-Net (ResNet-50)#822Top 1 Accuracy: 77.48%Number of params: 24.37MGFLOPs: 3.86
image-classification-on-imagenetECA-Net (MobileNetV2)#933Top 1 Accuracy: 72.56%Number of params: 3.34MGFLOPs: 0.320
object-detection-on-dsecECANet#8mAP: 25.7
object-detection-on-pku-ddd17-carECANet#5mAP50: 82.2

Rank counts only results with a code link.