Merge branch 'main' of hf.co:datasets/chengyenhsieh/TAO-Amodal

README.md

@@ -87,11 +87,19 @@ from huggingface_hub import snapshot_download
 snapshot_download(repo_id="chengyenhsieh/TAO-Amodal")
 ```
 
-2. Download all the video frames:
+2. Unzip all videos:
 
-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
+Modify `dataset_root` in [unzip_video.py](./unzip_video.py) and run:
+
+```bash
+python unzip_video.py
+```
+
+3. Download video frames through git (optional):
+
+You can either download the frames following the instructions [here](https://motchallenge.net/tao_download.php) (recommended) or modify our provided [script](./download_frames.sh) and run
 ```bash
-bash download_TAO.sh
+bash download_frames.sh
 ```
 
 

frames/train/YFCC100M.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a56ed50056ffa785775095c8ab34a394e6f24c54beeacfb2fef146c3a25aca39
+size 3410104357

frames/val/YFCC100M.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1eeec2655b8f5d2d9c98cfa1bd0d7e1e152a877e1a73517e0c818399eac38827
+size 6524282044