NVIDIA OmniDreams
Collection
3 items • Updated
video video 0.97 142 |
|---|
Curated single-view driving sequences for evaluating the
nvidia/alpadreams-dit world model.
data/
└── single_view/
├── <clip-id>/
| ├── <clip-id_...>.mp4 # ground truth video
│ ├── <clip-id_..._hdmap>.mp4 # HD-map rasterized conditioning video
│ ├── first_frame.png # RGB first frame, extracted from ground truth video
│ └── prompt.txt # text prompt
└── ...
| File | Type | Notes |
|---|---|---|
hdmap.mp4 |
video | rasterized HD-map conditioning, 720p / 30 fps |
first_frame.png |
image | RGB, matches the first frame of the target video |
prompt.txt |
text | free-form scenario description |
Small curated eval set. Not the training set.
Per §4.9 of the NVIDIA AV Dataset License Agreement, the dataset has been
anonymized with reasonable effort. Licensees must not attempt re-identification
or process license plates for tracking (§4.4). See LICENSE for the full terms,
including prohibitions on biometric processing and sensitive-attribute
inference (§4.5).
Scene geometry and demo scaffolding are inlined from the standalone
roaddreams project; primary author Yong He (yhe@nvidia.com). See
samples/roaddreams/NOTICE in the source distribution for the full attribution.