Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -14,19 +14,19 @@ Each datapoint has been trimmed using the [crop_video.py](https://github.com/che
|
|
| 14 |
Pose keypoints are predicted using [DWPose](https://github.com/IDEA-Research/DWPose?tab=readme-ov-file), and each sample is stored as a json file.
|
| 15 |
|
| 16 |
| Field | Type | Description |
|
| 17 |
-
|
| 18 |
-
| poses
|
| 19 |
-
|
|
| 20 |
-
| video_duration
|
| 21 |
-
| video_width
|
| 22 |
-
| video_height
|
| 23 |
-
| file_stem
|
| 24 |
-
| raw_text
|
| 25 |
-
| tokenized_text
|
| 26 |
-
| split
|
| 27 |
-
| yid
|
| 28 |
-
| start
|
| 29 |
-
| end
|
| 30 |
|
| 31 |
|
| 32 |
## References
|
|
|
|
| 14 |
Pose keypoints are predicted using [DWPose](https://github.com/IDEA-Research/DWPose?tab=readme-ov-file), and each sample is stored as a json file.
|
| 15 |
|
| 16 |
| Field | Type | Description |
|
| 17 |
+
|-------------------------|--------------|---------------------------------------------------------|
|
| 18 |
+
| poses | list | Pose data inferred from the video frames by DWPose |
|
| 19 |
+
| video_frames_per_second | float | Frame rate of the video |
|
| 20 |
+
| video_duration | float | Duration of the video in seconds |
|
| 21 |
+
| video_width | int | Width of the video |
|
| 22 |
+
| video_height | int | Height of the video |
|
| 23 |
+
| file_stem | str | File name stem (without extension) |
|
| 24 |
+
| raw_text | str | Raw textual description |
|
| 25 |
+
| tokenized_text | str | Tokenized version of the raw text |
|
| 26 |
+
| split | str | Dataset split (e.g., train/val/test) |
|
| 27 |
+
| yid | str | Youtube video ID |
|
| 28 |
+
| start | str | Start timestamp of the Youtube video segment |
|
| 29 |
+
| end | str | End timestamp of the Youtube video segment |
|
| 30 |
|
| 31 |
|
| 32 |
## References
|