ATLAS-WDS is a dataset for training models on wave directional spectra. It contains records with a 47x72 energy matrix flattened into a 3384-dimensional float32 array and corresponding skew-Gaussian anchor parameters. The dataset was created by author wuff-mann and was last updated on Hugging Face in May 2026.
Use Cases
- Training models to reconstruct wave energy spectra based on compressed anchor parameters.
- Developing algorithms for wave parameter estimation from spectral data.
- Benchmarking compressed sensing techniques for oceanographic remote sensing data.
Strengths
- Each record contains a precisely defined 47x72 energy matrix.
- Data includes corresponding skew-Gaussian anchor parameters for each energy matrix.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
Provenance
- Source
- wuff-mann on Hugging Face
- Freshness
- Last updated 2026-05-24 15:45:09; freshness should be verified.