Update README
Browse files
README.md
CHANGED
|
@@ -1,3 +1,85 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: mit
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
---
|
| 8 |
+
license: mit
|
| 9 |
+
language:
|
| 10 |
+
- en
|
| 11 |
+
metrics:
|
| 12 |
+
- recall
|
| 13 |
+
base_model:
|
| 14 |
+
- Qwen/Qwen2-VL-2B-Instruct
|
| 15 |
+
library_name: transformers == 4.45.2
|
| 16 |
+
---
|
| 17 |
+
|
| 18 |
+
<h1 align="center">Vis-IR: Unifying Search With Visualized Information Retrieval</h1>
|
| 19 |
+
|
| 20 |
+
<p align="center">
|
| 21 |
+
<a href="https://arxiv.org/abs/2502.11431">
|
| 22 |
+
<img alt="Build" src="http://img.shields.io/badge/arXiv-2502.11431-B31B1B.svg">
|
| 23 |
+
</a>
|
| 24 |
+
<a href="https://github.com/VectorSpaceLab/Vis-IR">
|
| 25 |
+
<img alt="Build" src="https://img.shields.io/badge/Github-Code-blue">
|
| 26 |
+
</a>
|
| 27 |
+
<a href="https://huggingface.co/datasets/marsh123/VIRA/">
|
| 28 |
+
<img alt="Build" src="https://img.shields.io/badge/🤗 Datasets-VIRA-yellow">
|
| 29 |
+
</a>
|
| 30 |
+
<a href="">
|
| 31 |
+
<img alt="Build" src="https://img.shields.io/badge/🤗 Datasets-MVRB-yellow">
|
| 32 |
+
</a>
|
| 33 |
+
<!-- <a href="">
|
| 34 |
+
<img alt="Build" src="https://img.shields.io/badge/🤗 Model-UniSE CLIP-yellow">
|
| 35 |
+
</a> -->
|
| 36 |
+
<a href="https://huggingface.co/marsh123/UniSE">
|
| 37 |
+
<img alt="Build" src="https://img.shields.io/badge/🤗 Model-UniSE MLLM-yellow">
|
| 38 |
+
</a>
|
| 39 |
+
|
| 40 |
+
</p>
|
| 41 |
+
|
| 42 |
+
## Overview
|
| 43 |
+
|
| 44 |
+
**VIRA** (Vis-IR Aggregation), a large-scale dataset comprising a vast collection of screenshots from diverse sources, carefully curated into captioned and questionanswer formats.
|
| 45 |
+
|
| 46 |
+
## Statistics
|
| 47 |
+
|
| 48 |
+
There are three types of data in VIRA: caption data, query-to-screenshot (q2s) data, and screenshot+query-to-screenshot (sq2s) data. The table below provides a detailed breakdown of the data counts for each domain and type.
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
## Organization Structure
|
| 54 |
+
|
| 55 |
+
The dataset is organized in the following structure:
|
| 56 |
+
|
| 57 |
+
```tree
|
| 58 |
+
Domain/
|
| 59 |
+
├── caption.jsonl: a screenshot image path and its corresponding caption
|
| 60 |
+
├── q2s.jsonl: a query, a positive screenshot and eight negative screenshots
|
| 61 |
+
├── sq2s.jsonl: a query, a query screenshot, a positive screenshot and eight negative screenshots
|
| 62 |
+
└── images/
|
| 63 |
+
├── image1.jpg
|
| 64 |
+
├── image2.jpg
|
| 65 |
+
...
|
| 66 |
+
```
|
| 67 |
+
_Due to the large number of images, uploading all of them takes time. Currently, the upload is not yet complete, and we will continue the process._
|
| 68 |
+
|
| 69 |
+
## License
|
| 70 |
+
VIRA is licensed under the [MIT License](LICENSE).
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
## Citation
|
| 74 |
+
If you find this dataset useful, please cite:
|
| 75 |
+
|
| 76 |
+
```
|
| 77 |
+
@article{liu2025any,
|
| 78 |
+
title={Any Information Is Just Worth One Single Screenshot: Unifying Search With Visualized Information Retrieval},
|
| 79 |
+
author={Liu, Ze and Liang, Zhengyang and Zhou, Junjie and Liu, Zheng and Lian, Defu},
|
| 80 |
+
journal={arXiv preprint arXiv:2502.11431},
|
| 81 |
+
year={2025}
|
| 82 |
+
}
|
| 83 |
+
```
|
| 84 |
+
|
| 85 |
+
|