Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- Wan-AI/Wan2.2-TI2V-5B
|
| 4 |
+
tags:
|
| 5 |
+
- perspective-to-360
|
| 6 |
+
extra gated eu disallowed: true
|
| 7 |
+
language:
|
| 8 |
+
- en
|
| 9 |
+
- zh
|
| 10 |
+
pipeline_tag: video-to-video
|
| 11 |
+
---
|
| 12 |
+
## CubeComposer Model
|
| 13 |
+
|
| 14 |
+
<p align="center"> <b> CubeComposer: Spatio-Temporal Autoregressive 4K 360° Video Generation from Perspective Video </b> </p>
|
| 15 |
+
|
| 16 |
+
<p align="center"> Lingen Li, Guangzhi Wang, Xiaoyu Li, Zhaoyang Zhang, Qi Dou, Jinwei Gu, Tianfan Xue, Ying Shan </p>
|
| 17 |
+
|
| 18 |
+
<p align="center"> CVPR 2026 </p>
|
| 19 |
+
|
| 20 |
+
**TL;DR**: Generate one cubemap face per time window with an effective and efficient context mechanism. Then, perspective video becomes 4K 360° without the memory blow‑up or the low‑res‑then‑upscale.
|
| 21 |
+
|
| 22 |
+
For more details, please visit our [project page](https://lg-li.github.io/project/cubecomposer/).
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
|