Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
|
@@ -11,34 +11,35 @@ tags:
|
|
| 11 |
- video-generation
|
| 12 |
- reasoning
|
| 13 |
---
|
|
|
|
| 14 |
<div align="center">
|
| 15 |
|
| 16 |
### We bet on a future where video reasoning is the next fundamental intelligence paradigm.
|
| 17 |
|
| 18 |
-
[π Website](https://video-reason.com)
|
| 19 |
|
| 20 |
</div>
|
| 21 |
|
| 22 |
-
---
|
| 23 |
## About Us
|
|
|
|
| 24 |
**Video-Reason** is a research initiative dedicated to advancing **video reasoning** as the next foundational intelligence paradigm β where spatiotemporal, embodied world experiences are more naturally captured than through text alone. We build large-scale datasets, benchmarks, and models to systematically study and scale video reasoning capabilities.
|
|
|
|
| 25 |
## VBVR: A Very Big Video Reasoning Suite
|
|
|
|
| 26 |
Our flagship project, **VBVR (Very Big Video Reasoning)**, introduces an unprecedentedly large-scale resource for video reasoning research:
|
|
|
|
| 27 |
- **200** curated reasoning tasks across 5 domains: **Perception**, **Abstraction**, **Spatiality**, **Transformation**, and **Knowledge**
|
| 28 |
- **1,000,000+** video clips β approximately three orders of magnitude larger than existing datasets
|
| 29 |
- **Verifiable evaluation** via rule-based, human-aligned scorers (no model-based judging)
|
| 30 |
- **Early signs of emergent generalization** to unseen reasoning tasks through large-scale scaling studies
|
| 31 |
-
---
|
| 32 |
|
| 33 |
## Links
|
| 34 |
|
| 35 |
-
- **Website**:
|
| 36 |
-
- **Paper**:
|
| 37 |
-
- **Slack**:
|
| 38 |
-
- **
|
| 39 |
-
- **Contact**:
|
| 40 |
-
|
| 41 |
-
---
|
| 42 |
|
| 43 |
## Citation
|
| 44 |
|
|
@@ -65,5 +66,4 @@ If you use VBVR in your research, please cite:
|
|
| 65 |
journal = {arXiv preprint arXiv:2602.20159},
|
| 66 |
year = {2026},
|
| 67 |
url = {https://arxiv.org/abs/2602.20159}
|
| 68 |
-
}
|
| 69 |
-
```
|
|
|
|
| 11 |
- video-generation
|
| 12 |
- reasoning
|
| 13 |
---
|
| 14 |
+
|
| 15 |
<div align="center">
|
| 16 |
|
| 17 |
### We bet on a future where video reasoning is the next fundamental intelligence paradigm.
|
| 18 |
|
| 19 |
+
[π Website](https://video-reason.com) Β· [π» GitHub](https://github.com/VBVR-DataFactory)
|
| 20 |
|
| 21 |
</div>
|
| 22 |
|
|
|
|
| 23 |
## About Us
|
| 24 |
+
|
| 25 |
**Video-Reason** is a research initiative dedicated to advancing **video reasoning** as the next foundational intelligence paradigm β where spatiotemporal, embodied world experiences are more naturally captured than through text alone. We build large-scale datasets, benchmarks, and models to systematically study and scale video reasoning capabilities.
|
| 26 |
+
|
| 27 |
## VBVR: A Very Big Video Reasoning Suite
|
| 28 |
+
|
| 29 |
Our flagship project, **VBVR (Very Big Video Reasoning)**, introduces an unprecedentedly large-scale resource for video reasoning research:
|
| 30 |
+
|
| 31 |
- **200** curated reasoning tasks across 5 domains: **Perception**, **Abstraction**, **Spatiality**, **Transformation**, and **Knowledge**
|
| 32 |
- **1,000,000+** video clips β approximately three orders of magnitude larger than existing datasets
|
| 33 |
- **Verifiable evaluation** via rule-based, human-aligned scorers (no model-based judging)
|
| 34 |
- **Early signs of emergent generalization** to unseen reasoning tasks through large-scale scaling studies
|
|
|
|
| 35 |
|
| 36 |
## Links
|
| 37 |
|
| 38 |
+
- **Website**: https://video-reason.com
|
| 39 |
+
- **Paper**: https://arxiv.org/abs/2602.20159
|
| 40 |
+
- **Slack**: https://join.slack.com/t/video-reason/shared_invite/zt-3qqf23icm-UC29fatWWYsIuzRNBR1lgg
|
| 41 |
+
- **Hugging Face**: https://huggingface.co/Video-Reason
|
| 42 |
+
- **Contact**: hokinxqdeng@gmail.com
|
|
|
|
|
|
|
| 43 |
|
| 44 |
## Citation
|
| 45 |
|
|
|
|
| 66 |
journal = {arXiv preprint arXiv:2602.20159},
|
| 67 |
year = {2026},
|
| 68 |
url = {https://arxiv.org/abs/2602.20159}
|
| 69 |
+
}
|
|
|