benchflow/frontierphysics-multi-object-visual-inverse-dynamics

FrontierPhysics: Multi-Object Visual Inverse Dynamics

Pinned copies of the upstream files used by multi-object-visual-inverse-dynamics in FrontierPhysics, hosted by BenchFlow. The task's image build and its verifier download these files from this repository instead of the upstream dataset.

This repository holds the ten two-object scenes the task uses: 2objs/{01,02,04,06,23,24,26,45,56,67}.zip. Each archive is byte-identical to the same path in Hanibel/MOSIV at revision bc1e42d3587420765965f6a27b92b4d9fe8bb987. MANIFEST.sha256 lists their SHA-256 digests in sha256sum -c format, and the task's verifier/reference_manifest.json pins each archive's size and digest.

The archives include the hidden evaluation frames and point clouds. FrontierPhysics uses them only in its verifier, and its image build exposes only the observed interval. The task's agent sandbox blocks this repository.

Oracle artifacts

oracle/ holds oracle-only packages from the task author, Yizhou Zhao. They are byte-for-byte copies of yizhouz/frontierphysics-oracle-artifacts at revision 9ab8539baad4f84f4a418c60d21ead7462950ab8. The task's oracle downloads the package for its assigned scene from a pinned revision of this repository and checks its SHA-256. A scene without a package is trained from scratch. The task's agent sandbox blocks this repository.

Path	Content
oracle/scene_01.zip	Scene 01: retained submission, model and trajectory; passed the task verifier in the author's check
oracle/scene_24.zip	Scene 24: retained submission, model and trajectory; passed the task verifier in the author's check
oracle/retained_current_results.json	The author's verifier record for both packages
oracle/MOSIV_training_pipeline_main.zip	Training and evaluation source of the task oracle
oracle/upstream/	The author's README, integration notes, licence notes, manifest and agent denylist

Scenes 02, 04, 06, 23, 26, 45, 56 and 67 have no package yet. oracle/MANIFEST.sha256 and oracle/provenance.json list the size and SHA-256 of every file and its upstream path. Source archives keep their original licences, including the GRAPHDECO notices for the Gaussian Splatting code.

Download

Pin a full commit revision; the FrontierPhysics task files record the one they use. The data files have not changed since commit 98974d4d2e76868b674605dfa6c9856806134181. This card and MANIFEST.sha256 were added after that commit.

from huggingface_hub import hf_hub_download

path = hf_hub_download(
    repo_id="benchflow/frontierphysics-multi-object-visual-inverse-dynamics",
    repo_type="dataset",
    revision="98974d4d2e76868b674605dfa6c9856806134181",
    filename="2objs/01.zip",
)

Attribution and licence

MOSIV is released under the MIT licence by its authors. Please cite:

@misc{liu2026mosivmultiobjectidentificationvideos,
      title={MOSIV: Multi-Object System Identification from Videos},
      author={Chunjiang Liu and Xiaoyuan Wang and Qingran Lin and Albert Xiao and Haoyu Chen and Shizheng Wen and Hao Zhang and Lu Qi and Ming-Hsuan Yang and Laszlo A. Jeni and Min Xu and Yizhou Zhao},
      year={2026},
      eprint={2603.06022},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2603.06022},
}

Browse the FrontierPhysics task datasets collection.