Skip to content

Dataset · subsurfacegen

3D Subsurface Velocity Models and Acoustic Wavefields

subsurfacegen/field-scale-dataset

Large geophysical benchmark of tens of thousands of 3D velocity volumes, 2D slices, and wavefields for seismic exploration.

What's inside

A large-scale benchmark ranging from 10,000 to 100,000 records in tabular and text formats (image-to-image task). Contains three-dimensional subsurface wave propagation velocity models (smoothed and depth-clipped to 619 samples) paired with 2D slices of corresponding acoustic wavefields and seismogram cubes from 64 sources. Data covers diverse geological conditions and five frequency bands from 3 to 25 Hz.

What you can do with it

  • Training neural networks for seismic inversion and wavefield prediction
  • Building geophysical models incorporating multiple signal sources
Tasks
Image → image
Size
10K–100K rows
Downloads this month
37K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("subsurfacegen/field-scale-dataset")