Update README.md

README.md

@@ -151,21 +151,21 @@ For all methods, the same human detection results are used (download link is pro
 
 | Methods | AP | AP .5 | AP .75 | AP (M) | AP (L) | AR | AR .5 | AR .75 | AR (M) | AR (L) | Download |
 |:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|
-| 256x192_resnet50<br>(this repo) | 70.4 | 88.6 | 77.8 | 67.0 | 76.9 | 76.2 | 93.0 | 83.0 | 71.9 | 82.4 | [model](https://drive.google.com/file/d/1mEi23LxpQQyqul7GipFbzEp9cIJ9zABe/view?usp=sharing)<br>[pose](https://drive.google.com/file/d/1Mw209h8XueWhd39S_CcIpbbZl2Zx5mEO/view?usp=sharing) |
+| 256x192_resnet50<br>(this repo) | 70.4 | 88.6 | 77.8 | 67.0 | 76.9 | 76.2 | 93.0 | 83.0 | 71.9 | 82.4 | [model](https://1drv.ms/u/s!All7gdNh7XE5pEsqottg_ZjOgg2m?e=fdlpjw)<br>[pose](https://1drv.ms/u/s!All7gdNh7XE5pEqhTEBI7fF6zFjM?e=a5IdvO) |
 | 256x192_resnet50<br>([original repo](https://github.com/Microsoft/human-pose-estimation.pytorch)) | 70.3 | 88.8 | 77.8 | 67.0 | 76.7 | 76.1 | 93.0 | 82.9 | 71.8 | 82.3 | - | 
 
-* Human detection result on val2017 (55.3 AP on human class) and test-dev2017 (57.2 AP on human class) [[bbox](https://drive.google.com/drive/folders/1E29nVHg0cbEjl4jnKpCEUlAsQPkoqnCY?usp=sharing)]
+* Human detection result on val2017 (55.3 AP on human class) and test-dev2017 (57.2 AP on human class) [[bbox](https://1drv.ms/f/s!All7gdNh7XE5pER6LuS-nIzWafJv?e=Cx6U4u)]
 * Other human detection results on val2017 [[Detectron_MODEL_ZOO](https://github.com/facebookresearch/Detectron/blob/master/MODEL_ZOO.md)]
 
 ### Results on PoseTrack 2018 dataset
 The pre-trained model on COCO dataset is used for training on the PoseTrack dataset following [paper](https://arxiv.org/abs/1804.06208). After training model on the COCO dataset, I set `lr`, `lr_dec_epoch`, `end_epoch` in `config.py` to `5e-5`, `[150, 155]`, `160`, respectively. Then, run `python train.py --gpu $GPUS --continue`. The table below is APs on validation set.
 
 | Methods | Head | Shoulder | Elbow | Wrist | Hip | Knee | Ankle | Total | Download |
 |:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|
-| 256x192_resnet50<br>(bbox from detector) | 74.4 | 76.9 | 72.2 | 65.2 | 69.2 | 70.0 | 62.9 | 70.4 | [model](https://drive.google.com/file/d/1_gU8eO2NjT24jlqbOMWpLIDjHMyedAqN/view?usp=sharing)<br>[pose](https://drive.google.com/file/d/1iRIMvu8kFxyrY--DwJcuZfeovydqOZFr/view?usp=sharing) |
-| 256x192_resnet50<br>(bbox from GT) | 87.9 | 86.7 | 80.2 | 72.5 | 77.0 | 77.8 | 74.6 | 80.1 | [model](https://drive.google.com/file/d/1_gU8eO2NjT24jlqbOMWpLIDjHMyedAqN/view?usp=sharing)<br>[pose](https://drive.google.com/file/d/1Gc_YtBNttuKsPJ4Pccrckio8OkOzKiQX/view?usp=sharing) |
+| 256x192_resnet50<br>(bbox from detector) | 74.4 | 76.9 | 72.2 | 65.2 | 69.2 | 70.0 | 62.9 | 70.4 | [model](https://1drv.ms/u/s!All7gdNh7XE5pD3q4Ywx9ELxLkDM?e=p0GD5z)<br>[pose](https://1drv.ms/u/s!All7gdNh7XE5pE13j2UbYJ6UDTaU?e=cIVNBP) |
+| 256x192_resnet50<br>(bbox from GT) | 87.9 | 86.7 | 80.2 | 72.5 | 77.0 | 77.8 | 74.6 | 80.1 | [model](https://1drv.ms/u/s!All7gdNh7XE5pD3q4Ywx9ELxLkDM?e=p0GD5z)<br>[pose](https://1drv.ms/u/s!All7gdNh7XE5pExciR53kCbdrWrk?e=BTn5WH) |
 
-* Human detection result on validation set [[bbox](https://drive.google.com/file/d/1bFx4Jd8b0iYADaA85BIUygKsaDZDP1i5/view?usp=sharing)]
+* Human detection result on validation set [[bbox](https://1drv.ms/u/s!All7gdNh7XE5pE5HD-9Hn7hQUGOX?e=5LOoht)]
 
 ## Troubleshooting
 1. Add graph.finalize when your machine takes more memory as training goes on. [[issue](https://github.com/mks0601/TF-SimpleHumanPose/issues/9)]