Update README.md

README.md

@@ -74,17 +74,17 @@ Our dataset encompasses 880 categories, aimed at assessing the occlusion reasoni
 through the paradigm of Tracking Any Object with Amodal perception (TAO-Amodal).
 
 ### Dataset Download
-1. Download all the annotations.
+1. Download with git:
 ```bash
 git lfs install
 git clone git@hf.co:datasets/chengyenhsieh/TAO-Amodal
 ```
 
-2. Download all the video frames:
+2. Download with [`python`](https://huggingface.co/docs/huggingface_hub/guides/download#download-files-from-the-hub):
 
-You can either download the frames following the instructions [here](https://motchallenge.net/tao_download.php) (recommended) or modify our provided [script](./download_TAO.sh) and run
-```bash
-bash download_TAO.sh
+```python
+from huggingface_hub import snapshot_download
+snapshot_download(repo_id="chengyenhsieh/TAO-Amodal")
 ```