Dataset · Hugging Face

ziqima/TrajectoryBench

Name: TrajectoryBench
Creator: ziqima
License: mit
Keywords: task_categories:image-to-video, license:mit, size_categories:1K<n<10K, format:json, modality:image, modality:text, library:datasets, library:pandas, library:polars, library:mlcroissant, arxiv:2610.03636, region:us, world-models, video-generation, benchmark

TrajectoryBench TrajectoryBench is a benchmark for long-horizon, camera-controlled video generation with complex trajectories. Each of its 2000 samples pairs a first frame with a camera trajectory and a text prompt. The model generates a video that starts from the frame and follows the trajectory, and is scored on 3D consistency, camera following, and video quality. Paper: LoGo: Local-Global Rewards for Consistent Long-Horizon Video Generation Contents… See the full description on the dataset page: https://huggingface.co/datasets/ziqima/TrajectoryBench.

View on Hugging FaceReport an issue

Author / organization	ziqima
License (as declared)	mit
Downloads	136 (as reported by Hugging Face)
Likes	0
Created	25 September 2026
Last updated at source	5 October 2026

world-modelsvideo-generationbenchmark

Embed this entry

<iframe src="https://silytics.si/embed/item/00928670-7a6f-4ca2-aa4a-1afd734be8b6" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

More from ziqima

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 16 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.