Dataset Viewer
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
Transient CH₄ concentration fields
Simulation data used by Laplace autoencoders: transient CH₄ concentration fields U(t)
as a function of physical parameters θ = (k, A, C). The simulations were generated with
OpenFOAM.
- Code: github.com/keyvanatt/laplace-autoencoders
- Checkpoints: keyvanatt/laplace-autoencoders-checkpoints
Files
| File | Shape | Content |
|---|---|---|
doe.npy |
(225,) structured (k, A, C) |
design of experiments: 15 values of k × 15 values of C, A = 0 |
split.npz |
train_idx (6480,), test_idx (1620,) |
train/test indices into the rotated dataset (8100 = 225 × 36) |
CH4.npy |
(225, 150, 200, 200) |
raw CH₄ fields: simulation × time step × grid |
ch4_rotated.npy (8100, 150, 128, 128) and doe_rotated.npy (8100,) are not stored here.
Regenerate them from CH4.npy and doe.npy:
hf download keyvanatt/laplace-autoencoders-dataset --repo-type dataset --local-dir dataset
PYTHONPATH=src python -m laplace_surrogate.utils.rotate --input_dir dataset --output_dir dataset
The rotation step produces 36 uniformly spaced rotations of each simulation, each center-cropped to 128 × 128.
The angle is stored in the A field of doe_rotated.npy.
- Downloads last month
- 89