Datasets:
Modalities:
Image
Formats:
imagefolder
Size:
< 1K
Tags:
onditional-video-generation
video-generation
diffusion-models
synthetic-video
3d-simulation
diffusion-training
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -19,7 +19,37 @@ tags:
|
|
| 19 |
- waypoint-conditioning
|
| 20 |
---
|
| 21 |
|
| 22 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
|
| 24 |
**This dataset was generated by the UNDERWORLD app by webXOS**
|
| 25 |
|
|
@@ -43,7 +73,8 @@ This dataset is optimized for low-end device simulation and micro-scale spatial
|
|
| 43 |
|
| 44 |
# Short dataset description (one-liner for card):
|
| 45 |
|
| 46 |
-
Synthetic 226-frame 1280×562 PNG sequence + metadata for training diffusion models in OVERWORLD-style environments
|
|
|
|
| 47 |
|
| 48 |
# Use cases:
|
| 49 |
|
|
@@ -55,5 +86,18 @@ Benchmarking lightweight video diffusion on constrained hardware.
|
|
| 55 |
# Integration with OVERWORLD (huggingface.co/overworld):
|
| 56 |
|
| 57 |
UNDERWORLD_Dataset_v1 is the low-end / micro counterpart to OVERWORLD-style datasets on Hugging Face.
|
| 58 |
-
It supplies short, lightweight, pose-controlled PNG sequences explicitly designed for training the same class
|
| 59 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
- waypoint-conditioning
|
| 20 |
---
|
| 21 |
|
| 22 |
+
[](https://webxos.netlify.app)
|
| 23 |
+
[](https://github.com/webxos/webxos)
|
| 24 |
+
[](https://huggingface.co/webxos)
|
| 25 |
+
[](https://x.com/webxos)
|
| 26 |
+
|
| 27 |
+
<div style="
|
| 28 |
+
background: #00FF00;
|
| 29 |
+
border-left: 4px solid #00FF00;
|
| 30 |
+
padding: 1.5rem;
|
| 31 |
+
margin: 2rem 0;
|
| 32 |
+
font-family: 'Fira Code', 'Courier New', monospace;
|
| 33 |
+
color: #00FF00;
|
| 34 |
+
border-radius: 0 8px 8px 0;
|
| 35 |
+
">
|
| 36 |
+
<pre style="
|
| 37 |
+
font-size: 12px;
|
| 38 |
+
line-height: 1.2;
|
| 39 |
+
margin: 0;
|
| 40 |
+
overflow-x: auto;
|
| 41 |
+
color: #00FF00;
|
| 42 |
+
">
|
| 43 |
+
_ _ _ _______ ___________ _ _ ___________ _ ______
|
| 44 |
+
| | | | \ | | _ \ ___| ___ \ | | || _ | ___ \ | | _ \
|
| 45 |
+
| | | | \| | | | | |__ | |_/ / | | || | | | |_/ / | | | | |
|
| 46 |
+
| | | | . ` | | | | __|| /| |/\| || | | | /| | | | | |
|
| 47 |
+
| |_| | |\ | |/ /| |___| |\ \\ /\ /\ \_/ / |\ \| |___| |/ /
|
| 48 |
+
\___/\_| \_/___/ \____/\_| \_|\/ \/ \___/\_| \_\_____/___/
|
| 49 |
+
|
| 50 |
+
</div>
|
| 51 |
+
|
| 52 |
+
# UNDERWORLD Dataset v1
|
| 53 |
|
| 54 |
**This dataset was generated by the UNDERWORLD app by webXOS**
|
| 55 |
|
|
|
|
| 73 |
|
| 74 |
# Short dataset description (one-liner for card):
|
| 75 |
|
| 76 |
+
Synthetic 226-frame 1280×562 PNG sequence + metadata for training diffusion models in OVERWORLD-style environments
|
| 77 |
+
with camera pose and temporal controls; optimized for low-end / micro-scale 3D simulation.
|
| 78 |
|
| 79 |
# Use cases:
|
| 80 |
|
|
|
|
| 86 |
# Integration with OVERWORLD (huggingface.co/overworld):
|
| 87 |
|
| 88 |
UNDERWORLD_Dataset_v1 is the low-end / micro counterpart to OVERWORLD-style datasets on Hugging Face.
|
| 89 |
+
It supplies short, lightweight, pose-controlled PNG sequences explicitly designed for training the same class
|
| 90 |
+
of diffusion models (OVERWORLD-style = controllable 3D world generation via camera waypoints + temporal signals).
|
| 91 |
+
Use it to pre-train or distill models before scaling to larger OVERWORLD-style data → enables faster iteration
|
| 92 |
+
on small hardware while preserving compatible conditioning format.
|
| 93 |
+
|
| 94 |
+
# Source
|
| 95 |
+
|
| 96 |
+
webXOS
|
| 97 |
+
webxos.netlify.app
|
| 98 |
+
github.com/webxos
|
| 99 |
+
huggingface.co/webxos
|
| 100 |
+
|
| 101 |
+
# License:
|
| 102 |
+
|
| 103 |
+
MIT
|