Link model card to Arxiv paper and update usage snippet

This PR updates the model card by:
- Linking it to the published paper on Hugging Face (https://huggingface.co/papers/2607.05247).
- Updating the quickstart code snippet to reflect the correct module import `lingbot_vision` and usage as documented in the GitHub repository.
- Correcting the BibTeX citation template to match the repository's updated citation format.

README.md

@@ -1,7 +1,11 @@
 ---
-license: apache-2.0
+datasets:
+- custom
 language:
 - en
+library_name: pytorch
+license: apache-2.0
+pipeline_tag: image-feature-extraction
 tags:
 - computer-vision
 - self-supervised-learning
@@ -11,10 +15,6 @@ tags:
 - depth-estimation
 - semantic-segmentation
 - pytorch
-datasets:
-- custom
-library_name: pytorch
-pipeline_tag: image-feature-extraction
 ---
 
 # LingBot-Vision
@@ -46,7 +46,7 @@ Each checkpoint contains backbone weights only. Training-time heads, optimizer s
 
 - **Repository:** https://github.com/robbyant/lingbot-vision
 - **Project Page:** https://technology.robbyant.com/lingbot-vision
-- **Technical Report:** coming soon
+- **Paper:** [Vision Pretraining for Dense Spatial Perception](https://huggingface.co/papers/2607.05247)
 
 ### Related Models
 
@@ -90,7 +90,7 @@ Load a pretrained backbone:
 ```python
 import torch
 
-from lbot_vision_infer import load_pretrained_backbone
+from lingbot_vision import load_pretrained_backbone, extract_patch_tokens, load_image
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 dtype = torch.bfloat16 if device == "cuda" else torch.float32
@@ -101,7 +101,15 @@ backbone, embed_dim = load_pretrained_backbone(
     dtype=dtype,
 )
 
-print(backbone.patch_size, embed_dim)
+img_norm, _, _ = load_image(
+    "examples/example.png",
+    size=512,
+    patch_size=backbone.patch_size,
+    mode="square",
+)
+patch_tokens, patch_grid = extract_patch_tokens(backbone, img_norm, device, dtype)
+
+print(patch_tokens.shape, patch_grid, embed_dim)
 ```
 
 The `variant` argument can be `giant`, `large`, `base`, or `small`. You can also pass an explicit Hugging Face model repo or a local directory to `load_pretrained_backbone`.
@@ -129,6 +137,7 @@ The `variant` argument can be `giant`, `large`, `base`, or `small`. You can also
 @article{lingbot-vision2026,
   title={Vision Pretraining for Dense Spatial Perception},
   author={Fu, Zelin and Tan, Bin and Sun, Changjiang and Liu, Shaohui and Zheng, Kecheng and Xu, Yinghao and Zhu, Xing and Shen, Yujun and Xue, Nan},
+  journal={arXiv preprint arXiv:2607.05247},
   year={2026}
 }
 ```
@@ -136,4 +145,4 @@ The `variant` argument can be `giant`, `large`, `base`, or `small`. You can also
 ## Model Card Contact
 
 - **Issues:** https://github.com/robbyant/lingbot-vision/issues
-- **Email:** fuzelin.fzl@antgroup.com, xuenan.xue@antgroup.com
+- **Email:** fuzelin.fzl@antgroup.com, xuenan.xue@antgroup.com