Update README.md
Browse files
README.md
CHANGED
|
@@ -19,6 +19,11 @@ ViewBench is a dataset for camera-conditioned long-horizon video world models. I
|
|
| 19 |
|
| 20 |
This dataset accompanies the paper [Consistent Video World Model With Geometry-Aware Rotary Position Embedding](https://arxiv.org/abs/2602.07854).
|
| 21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
## Highlights
|
| 23 |
|
| 24 |
- Complete yaw, pitch, and roll coverage for controlled camera motion.
|
|
|
|
| 19 |
|
| 20 |
This dataset accompanies the paper [Consistent Video World Model With Geometry-Aware Rotary Position Embedding](https://arxiv.org/abs/2602.07854).
|
| 21 |
|
| 22 |
+
Dataset mirrors:
|
| 23 |
+
|
| 24 |
+
- ModelScope: https://modelscope.cn/datasets/JEdward/viewbench-dataset
|
| 25 |
+
- Hugging Face: https://huggingface.co/datasets/JEdward/viewbench-dataset
|
| 26 |
+
|
| 27 |
## Highlights
|
| 28 |
|
| 29 |
- Complete yaw, pitch, and roll coverage for controlled camera motion.
|