Sign inGet started
Dataset · Hugging Face

INSAIT-Institute/TRAVL

TRAVL: Trajectory-Aware Video-Language Finetuning Dataset TRAVL is a finetuning dataset of short real vs implausible videos with targeted Q/A annotations probing physical plausibility (e.g., object permanence, collisions, shadows, trajectory consistency). Repository layout Data schema (current JSON) TRAVL_QA.json is a JSON list. Each item: { "id": "video_000", "video": "video_000.mp4", "conversations": [ {"from": "human", "value":… See the full description on the dataset page: https://huggingface.co/datasets/INSAIT-Institute/TRAVL.

Author / organizationINSAIT-Institute
License (as declared)mit
Downloads652 (as reported by Hugging Face)
Likes1
Created7 October 2025
Last updated at source7 October 2025
videovlmphysicsplausibilitytrajectoriesqafinetuning
Embed this entry
<iframe src="https://silytics.si/embed/item/f822346f-ff8a-48a9-b7e1-8092048de1e0" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Recent changes

  • 16 Oct 2026 Description: TRAVL: Trajectory-Aware Video-Language Finetuning Dataset TRAVL is a finetuning dataset of short real vs implausible videos with targeted Q/A annotations probing physical plausibility (e.g., object permanence, collisions, shadows, trajector → TRAVL: Trajectory-Aware Video-Language Finetuning Dataset TRAVL is a finetuning dataset of short real vs implausible videos with targeted Q/A annotations probing physical plausibility (e.g., object permanence, collisions, shadows, trajector

Changes detected by Silytics between refreshes of the source.

More from INSAIT-Institute

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 Information may be incomplete or out of date.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.