SparseWake/sparsewake
SparseWake SparseWake is a synthetic benchmark for sparse temporal hydrodynamic sensing. ICLR 2027 release The expanded release adds controlled multi-source mixtures and common-prior nearest-source tasks, with complete core data banks, reference checkpoints, a small review supplement, and reproduction code with a frozen wake-library input. Download release iclr2027-v1.0rc2 The version page lists the three archives, exact sizes, checksums, extraction instructions… See the full description on the dataset page: https://huggingface.co/datasets/SparseWake/sparsewake.
0275
1from __future__ import annotations2 3import numpy as np4 5 6def position_rmse(y_true: np.ndarray, y_pred: np.ndarray) -> float:7 dx = y_pred[:, 0] - y_true[:, 0]8 dy = y_pred[:, 1] - y_true[:, 1]9 return float(np.sqrt(np.mean(dx * dx) + np.mean(dy * dy)))10 11 12def theta_mae_deg(y_true: np.ndarray, y_pred: np.ndarray) -> float:13 diff = np.angle(np.exp(1j * (y_pred[:, 2] - y_true[:, 2])))14 return float(np.mean(np.abs(diff)) * 180.0 / np.pi)15 16 17def summarize(y_true: np.ndarray, y_pred: np.ndarray) -> dict[str, float]:18 out = {"position_rmse": position_rmse(y_true, y_pred)}19 if y_true.shape[1] >= 3 and y_pred.shape[1] >= 3:20 out["theta_rel_mae_deg"] = theta_mae_deg(y_true, y_pred)21 return out22 23