More top-level README updates

README.md

@@ -14,9 +14,9 @@ Patrick Labatut,
 Armand Joulin,
 Piotr Bojanowski
 
-[[`Paper`](https://arxiv.org/abs/2304.07193)] [[`Blog`](https://ai.facebook.com/blog/dino-v2-computer-vision-self-supervised-learning/)] [[`Demo`](https://dinov2.metademolab.com)] [[`BibTeX`](#citing-dinov2)]
+[[`Paper #1`](https://arxiv.org/abs/2304.07193)] [`Paper #2`](https://arxiv.org/abs/2309.16588)] [[`Blog`](https://ai.facebook.com/blog/dino-v2-computer-vision-self-supervised-learning/)] [[`Demo`](https://dinov2.metademolab.com)] [[`BibTeX`](#citing-dinov2)]
 
-PyTorch implementation and pretrained models for DINOv2. For details, see the paper: **[DINOv2: Learning Robust Visual Features without Supervision](https://arxiv.org/abs/2304.07193)**.
+PyTorch implementation and pretrained models for DINOv2. For details, see the papers: **[DINOv2: Learning Robust Visual Features without Supervision](https://arxiv.org/abs/2304.07193)** and **[Vision Transformers Need Registers](https://arxiv.org/abs/2309.16588)**.
 
 DINOv2 models produce high-performance visual features that can be directly employed with classifiers as simple as linear layers on a variety of computer vision tasks; these visual features are robust and perform well across domains without any requirement for fine-tuning. The models were pretrained on a dataset of 142 M images without using any labels or annotations.