Upload README.md

README.md

@@ -18,9 +18,9 @@ Humans intuitively perceive object shape and orientation from a single image, gu
 
 To facilitate this, we construct Objaverse-OA, a dataset of **14,832** orientation-aligned 3D models spanning **1,008 categories**. Leveraging Objaverse-OA, we fine-tune two representative 3D generative models based on multi-view diffusion and 3D variational autoencoder frameworks to produce aligned objects that generalize well to unseen objects across various categories.
 
-The dataset has two chunks, and all 3D models are stored in GLB format. The 3D models are named with their corresponding uids in the original Objaverse-LVIS dataset, and you can get their category labels by referring to the annotation in the Objaverse-LVIS dataset.
+The dataset has two chunks, and all 3D models are stored in GLB format. The 3D models are named with their corresponding uids in the original Objaverse-LVIS dataset, and you can get their category labels by referring to the annotation in the [Objaverse-LVIS](https://huggingface.co/datasets/allenai/objaverse) dataset.
 
-The license of our Objaverse-OA dataset is the same as the Objaverse dataset. For more details about the license, please refer to the declaration in the [Objaverse](https://objaverse.allenai.org/) dataset.
+The license of our Objaverse-OA dataset is the same as the Objaverse dataset. For more details about the license, please refer to the declaration in the [Objaverse](https://huggingface.co/datasets/allenai/objaverse) dataset.
 
 
 ## Citation