beno / README.md
Zhongning's picture
Upload folder using huggingface_hub
bf17186 verified
|
Raw
History Blame Contribute Delete
2.78 kB
metadata
license: other
tags:
  - BENO
  - CFD
  - neural operator
language:
  - en
  - zh

BENO

Dataset Description

The BENO dataset originates from the ICLR 2024 paper BENO: Boundary-Embedded Neural Operators for Elliptic PDEs and is designed for solving elliptic partial differential equations under complex boundary conditions. The data contains random boundary geometries with four, three, two, one, or no corners, all standardized to a 32 x 32 grid resolution.

Paper: BENO: Boundary-Embedded Neural Operators for Elliptic PDEs

Supported Tasks

Scenario Description
Elliptic PDE solving Predict the solution field from the boundary conditions and the right-hand side of the equation.
Boundary-condition research Compare solution performance under Dirichlet and Neumann boundary conditions.
Geometry generalization evaluation Evaluate model generalization across different random boundary shapes.
Neural operator research Provide standardized training and evaluation data for operator models such as BENO.

Dataset Format and Structure

The data is organized by boundary condition:

data/
  Dirichlet/
  Neumann/

Each boundary-condition category contains the following six configurations: N32_0c, N32_1c, N32_2c, N32_3c, N32_4c, and N32_mix. Each configuration contains 1,000 float64 samples:

File shape Description
BC_<prefix>_all.npy [1000, 128, 4] Boundary coordinates, boundary values, and boundary features.
RHS_<prefix>_all.npy [1000, 1024, 4] Coordinates, source terms, and cell states on a 32×32 grid.
SOL_<prefix>_all.npy [1000, 1024, 1] Elliptic PDE solution fields.

How to Use the Dataset

This dataset has been adapted for the OneScience-Group/BENO model. Download the dataset and model:

hf download --dataset OneScience-Group/beno --local-dir ./data

Official OneScience Information

Citation and License