Zhang, Wenqiang, Zilong Huang, Guozhong Luo, Tao Chen, Xinggang Wang, Wenyu Liu, Gang Yu,and Chunhua Shen. "TopFormer: Token Pyramid Transformer for Mobile Semantic Segmentation." In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 12083-12093. 2022.
| Model | Backbone | Resolution | Training Iters | mIoU | mIoU (flip) | mIoU (ms+flip) | Links |
|---|---|---|---|---|---|---|---|
| TopFormer-Base | topformer | 512x512 | 160000 | 38.28% | 38.59% | - | model | log | vdl |
| TopFormer-Small | topformer | 512x512 | 160000 | 35.60% | 35.83% | - | model | log | vdl |
| TopFormer-Tiny | topformer | 512x512 | 160000 | 32.49% | 32.75% | - | model | log | vdl |
Note that, the input resulution of TopFormer should be a multiple of 32.