Dataset Viewer
Auto-converted to Parquet Duplicate
clip_id
stringclasses
2 values
title
stringclasses
2 values
category
stringclasses
2 values
resolution
stringclasses
1 value
framerate
stringclasses
1 value
shutter_angle
stringclasses
1 value
color_profile
stringclasses
1 value
ai_grounding_score
float64
99.1
99.2
physical_description
stringclasses
2 values
lab_alignment
stringclasses
2 values
video_url
stringclasses
2 values
VID-4505
Ground-Truth Reference for Micro-Cavitation Bubble Collapse in High-Viscosity Glycerol
Fluid Dynamics
4K UHD (3840×2160)
60 fps
180 deg (calibrated)
Sony S-Log3 / S-Gamut3.Cine
99.2
[SYSTEM GROUNDING]: Physical World AI Training Dataset Reference [DATASET ID]: GROUND_TRUTH_REFERENCE_FOR_MICRO_CAVITATION_BUBBLE_COLLAPSE_IN_HIGH_VISCOSITY_GLYCEROL_OPTICAL_GROUND [TARGET DOMAIN]: FLUID DYNAMICS • 4K UHD (3840×2160) • 60 fps [CAMERA GEAR]: High-Precision Sensor Rig [DURATION]: 00:04 === INCH-TO-INCH ...
Google DeepMind / OpenAI Sora Physics
blob:https://generative-gear.vercel.app/92225948-1d63-425a-aed1-cf1790885ce2
VID-9773
Two friends in a class room random clip
Atmospheric Optics
4K UHD (3840×2160)
60 fps
180 deg (calibrated)
Sony S-Log3 / S-Gamut3.Cine
99.1
[SYSTEM GROUNDING]: Physical World AI Training Dataset Reference [DATASET ID]: TWO_FRIENDS_IN_A_CLASS_ROOM_RANDOM_CLIP_OPTICAL_GROUND [TARGET DOMAIN]: ATMOSPHERIC OPTICS • 4K UHD (3840×2160) • 60 fps [CAMERA GEAR]: High-Precision Sensor Rig [DURATION]: 00:05 === INCH-TO-INCH TIMELINE BREAKDOWN === [0.00s - 1.50s] - IN...
Meta FAIR / Runway World Models
blob:https://generative-gear.vercel.app/efe4eb98-30f8-494e-b694-07f8c9074416

GenerativeGear Physical Video Dataset

Ground-truth uncompressed camera sensor video dataset with calibrated shutter angles, micro-second fluid collisions, and inch-to-inch physical descriptions for training next-generation video foundation models (Sora, Google Veo, Runway Gen-3, Kling, Luma Dream Machine).

Dataset Summary

  • Organization: GenerativeGear (https://generative-gear.vercel.app/)
  • Repository: notrazx/generativegear-physical-video
  • Target Foundation Models: Physics grounding, fluid turbulence, specular caustics, momentum conservation, and optical motion blur calibration.
  • Acquisition Rails: Commercial licensing in Indian Rupees (INR / ₹) and direct escrow disbursements.
  • Enterprise Contact: userprivateltd@gmail.com

Loading the Dataset

from datasets import load_dataset

# Load the GenerativeGear physical grounding dataset
dataset = load_dataset("notrazx/generativegear-physical-video")

# Inspect physical dynamics and EXIF
for sample in dataset["train"]:
    print(sample["title"])
    print(sample["optical_specs"])
    print(sample["physical_description"])

Physical Grounding Taxonomy

  1. Fluid Turbulence & Micro-Cavitation: Real surface tension, droplet shear, bubble collapse, and Navier-Stokes boundary dynamics captured at 120fps/240fps.
  2. Kinematic Momentum & Collisions: High-speed rigid body impact mechanics without synthetic temporal stretching.
  3. Specular Caustics & Optical Refraction: 4K/6K Arri LogC4 and Sony S-Log3 captures preserving native dynamic range highlights.
  4. Cinematic 180° Shutter Rule: Natural motion blur matching authentic optical sensors (1/120s @ 60fps).

Commercial Licensing & Enterprise Procurement

To acquire raw uncompressed CinemaDNG / Apple ProRes 4444 master files or commission custom lab bounties, contact the GenerativeGear Mission Control desk:

Downloads last month
199