GTP-ViT: Efficient Vision Transformers via Graph-based Token Propagation

Open paper
Benchmark Model Rank Results
image-classification-on-imagenetGTP-ViT-B-Patch8/P20#194Top 1 Accuracy: 85.8%
image-classification-on-imagenetGTP-EVA-L/P8#231Top 1 Accuracy: 85.4%
image-classification-on-imagenetGTP-ViT-L/P8#389Top 1 Accuracy: 83.7%
image-classification-on-imagenetGTP-LV-ViT-M/P8#487Top 1 Accuracy: 82.8%GFLOPs: 8
image-classification-on-imagenetGTP-LV-ViT-S/P8#579Top 1 Accuracy: 81.9%GFLOPs: 4.8
image-classification-on-imagenetGTP-DeiT-B/P8#612Top 1 Accuracy: 81.5%GFLOPs: 13.1
image-classification-on-imagenetGTP-DeiT-S/P8#710Top 1 Accuracy: 79.5%GFLOPs: 3.4

Rank counts only results with a code link.