Update README.md

README.md

@@ -96,16 +96,17 @@ Reference: Jiang et al., **RTMPose: Real-Time Multi-Person Pose Estimation based
 ## Usage
 
 ```python
-from datasets import load_dataset
+from datasets import load_dataset, concatenate_datasets
 
 images = load_dataset("ryushinn/11k-Hands", split="train")
 ann = load_dataset("ryushinn/11k-Hands-BBox-Keypoint", split="train")
 
+ds = concatenate_datasets([images, ann], axis=1)
 idx = 0
-image = images[idx]["image"]
-bbox = ann[idx]["bboxes_xyxy"]
-keypoints_xy = ann[idx]["keypoints_xy"]
-keypoint_scores = ann[idx]["keypoint_scores"]
+image = ds[idx]["image"]
+bbox = ds[idx]["bboxes_xyxy"]
+keypoints_xy = ds[idx]["keypoints_xy"]
+keypoint_scores = ds[idx]["keypoint_scores"]
 ```
 
 ## Visualization previews