| semantic-segmentation-on-pascal-context | ResNeSt-269 | #13 | mIoU: 58.9 |
| semantic-segmentation-on-pascal-context | ResNeSt-200 | #15 | mIoU: 58.4 |
| semantic-segmentation-on-pascal-context | ResNeSt-101 | #19 | mIoU: 56.5 |
| semantic-segmentation-on-cityscapes-test | ResNeSt200 (Mapillary) | #14 | Mean IoU (class): 83.3% |
| semantic-segmentation-on-dada-seg | ResNeSt (ResNeSt-101) | #21 | mIoU: 19.99 |
| panoptic-segmentation-on-coco-minival | PanopticFPN+ResNeSt(single-scale) | #26 | PQ: 47.9PQst: 37.0PQth: 55.1 |
| semantic-segmentation-on-cityscapes-val | ResNeSt-200 | #29 | mIoU: 82.7 |
| instance-segmentation-on-coco-test-dev | ResNeSt101 | #49 | mask AP: 43% |
| instance-segmentation-on-coco-test-dev | ResNeSt-200 (multi-scale) | #105 | AP50: 70.2AP75: 51.5APS: 30.0APM: 49.6APL: 60.6 |
| instance-segmentation-on-coco-minival | ResNeSt-200 (multi-scale) | #53 | mask AP: 46.25 |
| instance-segmentation-on-coco-minival | ResNeSt-200-DCN (single-scale) | #65 | mask AP: 44.5 |
| instance-segmentation-on-coco-minival | ResNeSt-200 (single-scale) | #71 | mask AP: 44.21 |
| instance-segmentation-on-coco-minival | ResNeSt-101 (single-scale) | #81 | mask AP: 41.56 |
| semantic-segmentation-on-ade20k-val | ResNeSt-200 | #60 | mIoU: 48.36 |
| semantic-segmentation-on-ade20k-val | ResNeSt-269 | #62 | mIoU: 47.60 |
| semantic-segmentation-on-ade20k-val | ResNeSt-101 | #64 | mIoU: 46.91 |
| object-detection-on-coco-test-dev | ResNeSt-200 (Cascade R-CNN, DCN, multi-scale test; official repo) | #62 | box mAP: 53.3 |
| semantic-segmentation-on-ade20k | ResNeSt-200 | #162 | Validation mIoU: 48.36 |
| semantic-segmentation-on-ade20k | ResNeSt-269 | #177 | Validation mIoU: 47.60 |
| semantic-segmentation-on-ade20k | ResNeSt-101 | #184 | Validation mIoU: 46.91 |
| object-detection-on-coco-val2017 | ResNeSt-200 (Cascade R-CNN, DCN, tricks, 3x; official repo) | #198 | box AP: 50.91 |
| object-detection-on-coco-val2017 | ResNeSt-200 (Cascade R-CNN, tricks, 3x; official repo) | #206 | box AP: 50.54 |
| image-classification-on-imagenet | ResNeSt-269 | #330 | Top 1 Accuracy: 84.5%Number of params: 111M |
| image-classification-on-imagenet | ResNeSt-200 | #401 | Top 1 Accuracy: 83.9%Number of params: 70M |
| image-classification-on-imagenet | ResNeSt-101 | #504 | Top 1 Accuracy: 83.0%Number of params: 48M |
| image-classification-on-imagenet | ResNeSt-50 | #700 | Top 1 Accuracy: 81.13%Number of params: 27.5MGFLOPs: 5.39 |
| image-classification-on-imagenet | ResNeSt-50-fast | #729 | Top 1 Accuracy: 80.64%Number of params: 27.5MGFLOPs: 4.34 |