Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -15,6 +15,11 @@ size_categories:
|
|
| 15 |
|
| 16 |
# CRONOS-Benchmark dataset
|
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
CRONOS-Benchmark is a controlled, synthetic benchmark for evaluating **counterfactual physical consistency** in video world models. It tests whether generative video models correctly simulate three fundamental physical event types — **object falling**, **object collision**, and **object occlusion** — across diverse scenes, objects, and viewpoints.
|
| 19 |
|
| 20 |
Each sequence provides ground-truth RGB frames, depth maps, and segmentation masks, along with a 5-frame conditioning clip and a `metadata.json` file, enabling standardised evaluation of any video generation model that accepts image or video conditioning.
|
|
@@ -27,12 +32,6 @@ CRONOS-Benchmark dataset contains **675 sequences** rendered in a controlled sim
|
|
| 27 |
|
| 28 |
- **License:** Apache 2.0
|
| 29 |
|
| 30 |
-
### Dataset Sources
|
| 31 |
-
|
| 32 |
-
- **Project Page** [https://genintel.github.io/CRONOS/]
|
| 33 |
-
- **Repository:** [https://github.com/GenIntel/CRONOS-benchmark]
|
| 34 |
-
- **Paper:** [https://arxiv.org/abs/2605.23699]
|
| 35 |
-
|
| 36 |
|
| 37 |
## Dataset Structure
|
| 38 |
|
|
|
|
| 15 |
|
| 16 |
# CRONOS-Benchmark dataset
|
| 17 |
|
| 18 |
+
[](https://arxiv.org/abs/2605.23699)
|
| 19 |
+
[](https://genintel.github.io/CRONOS/)
|
| 20 |
+
[](https://huggingface.co/datasets/genintel/CRONOS-benchmark)
|
| 21 |
+
[](https://github.com/GenIntel/CRONOS-benchmark)
|
| 22 |
+
|
| 23 |
CRONOS-Benchmark is a controlled, synthetic benchmark for evaluating **counterfactual physical consistency** in video world models. It tests whether generative video models correctly simulate three fundamental physical event types — **object falling**, **object collision**, and **object occlusion** — across diverse scenes, objects, and viewpoints.
|
| 24 |
|
| 25 |
Each sequence provides ground-truth RGB frames, depth maps, and segmentation masks, along with a 5-frame conditioning clip and a `metadata.json` file, enabling standardised evaluation of any video generation model that accepts image or video conditioning.
|
|
|
|
| 32 |
|
| 33 |
- **License:** Apache 2.0
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
|
| 36 |
## Dataset Structure
|
| 37 |
|