Datasets:
You need to agree to share your Hugging Face information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
By requesting access, you agree to the VisionSupply Dataset Evaluation License v1.0 linked on this card.
Log in or Sign Up to review the conditions and access this dataset content.
VisionSupply.AI — Cinematic Video Datasets up to 8K
License cinematic source footage at resolutions up to 8K for AI training, fine-tuning, evaluation, and research. VisionSupply.AI curates custom datasets around the scenes, actions, camera movement, and metadata your project needs.
On Hugging Face: 14 downloadable preview clips totaling 4 minutes and 47 seconds, with clip-level descriptions and technical metadata. The previews are 1920 pixels wide. Commercial deliveries can include original footage at resolutions up to 8K.
Commercial delivery formats
Camera RAW and Apple ProRes are available on request for selected footage. Tell us your preferred source or delivery format, resolution, and frame-rate requirements. We confirm the native resolution, available codec or ProRes profile, and any color-space or grading requirements for the selected clips when quoting.
The Hugging Face preview package uses H.264 MP4 for easy playback. Commercial licensing can include higher-resolution source footage for workflows such as color grading, look development, and fine-detail video analysis.
What the footage offers
- Detail for fine-grained visual tasks. License high-resolution source footage when small objects, distant signage, or surface texture matter to your workflow.
- Variation in lighting and framing. The sample spans backlit street scenes, stage lighting, aerial wide shots, and extreme macro close-ups.
- Motion over time. Inspect camera motion, moving subjects, and occlusion in clips ranging from a forklift in a warehouse to dancers on a stage.
- Custom selection by brief. Request specific subjects, actions, settings, camera moves, and metadata fields for your intended use.
Request a larger sample or a licensing quote
We are preparing a commercial evaluation sample of 30 clips delivered in 4K. Contact us to discuss availability and evaluation terms.
Tell us what you’re building, which footage categories you need, and roughly how many clips or hours you’re looking for. We’ll confirm coverage, delivery options, and licensing terms for your intended use.
Email license@visionsupply.ai or visit visionsupply.ai.
Example AI workflows
- Color grading and look development. Inspect how color, contrast, and lighting vary across stage lighting, backlit streets, aerial footage, and macro imagery. The sample contains finished previews. Source format and paired grading references can be discussed for a custom delivery.
- Generative video and video editing. Inspect framing, camera movement, lighting styles, and subject motion when briefing a licensed dataset for a generation or editing workflow.
- Video captioning and temporal understanding. Explore clip descriptions and structured metadata for visible actions, subjects, camera movement, and on-screen text.
Footage provenance and rights
VisionSupply curates footage from a production library for licensing. For commercial orders, we confirm provenance, the scope of AI usage rights, and the availability and scope of model and property releases for the selected clips during licensing review. Contact us if your project needs clip-level documentation or specific release coverage.
Featured frames
More sample frames
The sample clips are silent H.264 MP4 previews at 1920 × 1012 or 1920 × 1080. The original source footage is available at resolutions up to 8K through commercial licensing.
Load the evaluation sample
After access is approved, authenticate with Hugging Face and install datasets and torchcodec to load and decode the previews.
from datasets import load_dataset
sample = load_dataset("visionsupply/cinematic-video-dataset", split="preview")
The preview/ directory contains 14 MP4 previews (visionsupply-001.mp4 through visionsupply-014.mp4) and metadata.jsonl. Each file_name points to the matching MP4. Title, scene description, shot scale, camera movement, lighting, and actions have their own fields. Additional annotations are stored in a structured object.
The resolution, frame rate, codec, and audio fields describe the delivered preview. Native resolution, camera and lens details, source format, and original audio availability are confirmed for selected footage when quoting a commercial delivery.
Annotations are provisional. Review them against the footage before using them as ground truth for model scoring.
Access and use
Approved teams may run internal model inference, generate embeddings, and calculate evaluation metrics under the VisionSupply Dataset Evaluation License v1.0.
- Downloads last month
- 308













