---
datasets:
- kitti
library_name: pytorch
license: apache-2.0
pipeline_tag: depth-estimation
tags:
- deltatok
- cvpr2026-highlight
---

# Depth Head — KITTI

Monocular depth estimation head trained on KITTI (RMSE: 2.79). Part of [A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens](https://huggingface.co/papers/2604.04913) (CVPR 2026 Highlight).

## Usage

Requires a frozen [DINOv3](https://github.com/facebookresearch/dinov3) ViT-B backbone. See the [DeltaTok GitHub repository](https://github.com/amazon-far/deltatok) for training and evaluation code.

## Acknowledgements

- [DINOv3](https://github.com/facebookresearch/dinov3)
- [KITTI](https://www.cvlibs.net/datasets/kitti/)

## Citation

```bibtex
@inproceedings{kerssies2026deltatok,
  title     = {A Frame is Worth One Token: Efficient Generative World Modeling with Delta Tokens},
  author    = {Kerssies, Tommie and Berton, Gabriele and He, Ju and Yu, Qihang and Ma, Wufei and de Geus, Daan and Dubbelman, Gijs and Chen, Liang-Chieh},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year      = {2026}
}
```