Add files using upload-large-folder tool
Browse files- README.md +18 -0
- config.json +10 -0
README.md
ADDED
|
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
pipeline_tag: image-to-3d
|
| 4 |
+
library_name: oneworld
|
| 5 |
+
tags:
|
| 6 |
+
- 3d-gaussian-splatting
|
| 7 |
+
- image-to-3d
|
| 8 |
+
- multi-view-generation
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# OneWorld
|
| 12 |
+
|
| 13 |
+
Four-step OneWorld checkpoint for image-conditioned 3D Gaussian scene generation.
|
| 14 |
+
It supports image-and-camera inference with optional text conditioning. Use the
|
| 15 |
+
inference code from `SensenGao/OneWorld`.
|
| 16 |
+
|
| 17 |
+
The package contains the distilled generator, the adapted RAE decoder, and latent
|
| 18 |
+
statistics. Pi3X and Wan2.1-T2V-1.3B are downloaded separately.
|
config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format": "oneworld-4step-v1",
|
| 3 |
+
"checkpoint": "checkpoint",
|
| 4 |
+
"rae": "rae.pt",
|
| 5 |
+
"stats": "latent_stats.pt",
|
| 6 |
+
"text_length": 256,
|
| 7 |
+
"views": 8,
|
| 8 |
+
"height": 224,
|
| 9 |
+
"width": 448
|
| 10 |
+
}
|