| image-classification-on-inaturalist | SpineNet-143 | #15 | Top 1 Accuracy: 63.6%Top 5 Accuracy: 84.8% |
| instance-segmentation-on-coco-test-dev | Mask R-CNN (SpineNet-190, 1536x1536) | #38 | mask AP: 46.1 |
| instance-segmentation-on-coco-minival | RetinaNet (SpineNet-190, 1536x1536) | #55 | mask AP: 46.1 |
| object-detection-on-coco-test-dev | RetinaNet (SpineNet-190, 1280x1280) | #70 | box mAP: 52.1AP50: 71.8AP75: 56.5APS: 35.4APM: 55APL: 63.6 |
| object-detection-on-coco-test-dev | RetinaNet (SpineNet-143, 1280x1280) | #82 | box mAP: 50.7AP50: 70.4AP75: 54.9APS: 33.6APM: 53.9APL: 62.1 |
| object-detection-on-coco-test-dev | RetinaNet (SpineNet-96, 1024x1024) | #97 | box mAP: 48.6AP50: 68.4AP75: 52.5APS: 32APM: 52.3APL: 62 |
| object-detection-on-coco-test-dev | RetinaNet (SpineNet-49, 896x896) | #114 | box mAP: 46.7AP50: 66.3AP75: 50.6APS: 29.1APM: 50.1APL: 61.7 |
| object-detection-on-coco-test-dev | RetinaNet (SpineNet-49, 640x640) | #131 | box mAP: 44.3AP50: 63.8AP75: 47.6APS: 25.9APM: 47.7APL: 61.1 |
| object-detection-on-coco-test-dev | SpineNet-49 (640, RetinaNet, single-scale) | #152 | box mAP: 42.8AP50: 62.3AP75: 46.1APS: 23.7APM: 45.2APL: 57.3 |
| object-detection-on-coco-test-dev | RetinaNet (SpineNet-49S, 640x640) | #164 | box mAP: 41.5AP50: 60.5AP75: 44.6APS: 23.3APM: 45APL: 58 |
| object-detection-on-coco-val2017 | Mask R-CNN (SpineNet-190, 1536x1536) | #173 | box AP: 52.2 |
| image-classification-on-imagenet | SpineNet-143 | #829 | Top 1 Accuracy: 79%Number of params: 60.5MGFLOPs: 9.1 |