Visual Attention Network

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetVAN-B6 (22K, 384res)#69Top 1 Accuracy: 87.8%Number of params: 200MGFLOPs: 114.3
image-classification-on-imagenetVAN-B5 (22K, 384res)#109Top 1 Accuracy: 87%GFLOPs: 50.6
image-classification-on-imagenetVAN-B6 (22K)#114Top 1 Accuracy: 86.9%Number of params: 200MGFLOPs: 38.9
image-classification-on-imagenetVAN-B4 (22K, 384res)#130Top 1 Accuracy: 86.6%Number of params: 60MGFLOPs: 35.9
image-classification-on-imagenetVAN-B5 (22K)#156Top 1 Accuracy: 86.3%Number of params: 90MGFLOPs: 17.2
image-classification-on-imagenetVAN-B4 (22K)#202Top 1 Accuracy: 85.7%GFLOPs: 12.2
image-classification-on-imagenetVAN-B2#483Top 1 Accuracy: 82.8%Number of params: 26.6MGFLOPs: 5
image-classification-on-imagenetVAN-B1#636Top 1 Accuracy: 81.1%Number of params: 13.9MGFLOPs: 2.5
image-classification-on-imagenetVAN-B0#889Top 1 Accuracy: 75.4%Number of params: 4.1MGFLOPs: 0.9
panoptic-segmentation-on-coco-minivalVisual Attention Network (VAN-B6 + Mask2Former)#10PQ: 58.2PQst: 48.2PQth: 64.8
semantic-segmentation-on-ade20kVAN-B6#54Validation mIoU: 54.7
semantic-segmentation-on-ade20kVAN-Large (HamNet)#116Validation mIoU: 50.2Params (M): 55
semantic-segmentation-on-ade20kVAN-Large#150Validation mIoU: 48.1Params (M): 49
semantic-segmentation-on-ade20kVAN-Base (Semantic-FPN)#168Validation mIoU: 46.7
semantic-segmentation-on-ade20kVAN-Small#205Validation mIoU: 42.9Params (M): 18
semantic-segmentation-on-ade20kVAN-Tiny#212Validation mIoU: 38.5Params (M): 8

Rank counts only results with a code link.