Update README.md
Browse files
README.md
CHANGED
|
@@ -115,7 +115,7 @@ computation to non-overlapping local windows while also allowing for cross-windo
|
|
| 115 |
Swin Transformer achieves strong performance on COCO object detection (`58.7 box AP` and `51.1 mask AP` on test-dev) and
|
| 116 |
ADE20K semantic segmentation (`53.5 mIoU` on val), surpassing previous models by a large margin.
|
| 117 |
|
| 118 |
-
 and
|
| 116 |
ADE20K semantic segmentation (`53.5 mIoU` on val), surpassing previous models by a large margin.
|
| 117 |
|
| 118 |
+
|
| 119 |
|
| 120 |
## Main Results on ImageNet with Pretrained Models
|
| 121 |
|