Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,54 @@
|
|
| 1 |
---
|
| 2 |
-
license:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
|
|
|
| 5 |
|
| 6 |
-
|
|
|
|
| 1 |
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
pipeline_tag: image generation
|
| 6 |
+
tags:
|
| 7 |
+
- mamba
|
| 8 |
+
- generative model
|
| 9 |
+
- stable diffusion
|
| 10 |
+
- stochastic interpolant
|
| 11 |
+
- zigma
|
| 12 |
+
- zigzag
|
| 13 |
+
|
| 14 |
---
|
| 15 |
+
# ZigMa: Zigzag Mamba Diffusion Model
|
| 16 |
+
|
| 17 |
+
This model represents the official checkpoint of the paper titled "ZigMa: Zigzag Mamba Diffusion Model".
|
| 18 |
+
|
| 19 |
+
[](https://https://taohu.me/project_zigma)
|
| 20 |
+
[](https://github.com/dongzhuoyao/zigma)
|
| 21 |
+
[](https://arxiv.orgg)
|
| 22 |
+
[](https://www.apache.org/licenses/LICENSE-2.0)
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
[Bingxin Ke](http://www.kebingxin.com/),
|
| 26 |
+
[Anton Obukhov](https://www.obukhov.ai/),
|
| 27 |
+
[Shengyu Huang](https://shengyuh.github.io/),
|
| 28 |
+
[Nando Metzger](https://nandometzger.github.io/),
|
| 29 |
+
[Rodrigo Caye Daudt](https://rcdaudt.github.io/),
|
| 30 |
+
[Konrad Schindler](https://scholar.google.com/citations?user=FZuNgqIAAAAJ&hl=en )
|
| 31 |
+
|
| 32 |
+
We present Marigold, a diffusion model and associated fine-tuning protocol for monocular depth estimation. Its core principle is to leverage the rich visual knowledge stored in modern generative image models. Our model, derived from Stable Diffusion and fine-tuned with synthetic data, can zero-shot transfer to unseen data, offering state-of-the-art monocular depth estimation results.
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
## 🎓 Citation
|
| 38 |
+
|
| 39 |
+
```bibtex
|
| 40 |
+
@InProceedings{ke2023repurposing,
|
| 41 |
+
title={Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation},
|
| 42 |
+
author={Bingxin Ke and Anton Obukhov and Shengyu Huang and Nando Metzger and Rodrigo Caye Daudt and Konrad Schindler},
|
| 43 |
+
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
|
| 44 |
+
year={2024}
|
| 45 |
+
}
|
| 46 |
+
```
|
| 47 |
+
|
| 48 |
+
## 🎫 License
|
| 49 |
+
|
| 50 |
+
This work is licensed under the Apache License, Version 2.0 (as defined in the [LICENSE](LICENSE.txt)).
|
| 51 |
|
| 52 |
+
By downloading and using the code and model you agree to the terms in the [LICENSE](LICENSE.txt).
|
| 53 |
|
| 54 |
+
[](https://www.apache.org/licenses/LICENSE-2.0)
|