AnteaWu/PDI-Dataset
PDI-Bench: Quantitative Video World Model Evaluation for Geometric-Consistency Project Page | Paper | GitHub PDI-Bench (Perspective Distortion Index) is a quantitative framework and dataset for auditing geometric coherence in generated videos. It evaluates whether video generation models produce physically plausible 3D structure and motion by computing residuals across three failure dimensions: Scale-Depth Alignment ($\epsilon_{scale}$): Measures whether object scale changes… See the full description on the dataset page: https://huggingface.co/datasets/AnteaWu/PDI-Dataset.
| Author / organization | AnteaWu |
|---|---|
| License (as declared) | other |
| Downloads | 238 (as reported by Hugging Face) |
| Likes | 2 |
| Created | 5 May 2026 |
| Last updated at source | 16 May 2026 |
Popularity, last 3 days
238 now (-59 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
Recent changes
- 16 Oct 2026 Description: PDI-Bench: Quantitative Video World Model Evaluation for Geometric-Consistency Project Page | Paper | GitHub PDI-Bench (Perspective Distortion Index) is a quantitative framework and dataset for auditing geometric coherence in generated vide → PDI-Bench: Quantitative Video World Model Evaluation for Geometric-Consistency Project Page | Paper | GitHub PDI-Bench (Perspective Distortion Index) is a quantitative framework and dataset for auditing geometric coherence in generated vide
Changes detected by Silytics between refreshes of the source.
Related papers
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.