Datasets:
Modalities:
Image
Formats:
imagefolder
Languages:
English
Size:
1K - 10K
Tags:
image
contextual-observer-grounding
spatial-reasoning
perspective-taking
vision-language-models
embodied-ai
License:
Download NOTICE from owl-owl/POVBench: direct link, hf CLI and curl.
- Browser
- Download file 1.28 kB
-
https://huggingface.co/datasets/owl-owl/POVBench/resolve/main/NOTICE
- Command line
-
hf download hf://datasets/owl-owl/POVBench/NOTICE
-
curl -L -o NOTICE https://huggingface.co/datasets/owl-owl/POVBench/resolve/main/NOTICE
1.28 kB
| POVBench | |
| Copyright (c) 2026 Mimo Shirasaka, Haochen Zhang, Yonatan Bisk | |
| This dataset contains material derived from the following Apache License 2.0 | |
| projects of the Allen Institute for AI. The rendered images and the scene | |
| metadata redistributed here remain subject to that license. | |
| -------------------------------------------------------------------------- | |
| AI2-THOR | |
| Copyright The Allen Institute for Artificial Intelligence | |
| Licensed under the Apache License, Version 2.0 | |
| https://github.com/allenai/ai2thor | |
| Kolve et al. "AI2-THOR: An Interactive 3D Environment for Visual AI." | |
| arXiv:1712.05474, 2017. | |
| -------------------------------------------------------------------------- | |
| ProcTHOR | |
| Copyright The Allen Institute for Artificial Intelligence | |
| Licensed under the Apache License, Version 2.0 | |
| https://github.com/allenai/procthor | |
| Deitke et al. "ProcTHOR: Large-Scale Embodied AI Using Procedural | |
| Generation." NeurIPS 2022. | |
| -------------------------------------------------------------------------- | |
| House layouts were procedurally generated with ProcTHOR and rendered with the | |
| AI2-THOR CloudRendering backend. Object category names and asset identifiers | |
| appearing in the JSON files (e.g. "ArmChair", "Floor_Lamp_8") originate from | |
| the AI2-THOR asset library. | |