coarse-grained-forecasting-benchmark's picture
Add anonymous reproduction code
41dd7e0 verified
Raw History Blame Contribute Delete
1.45 kB
# 2-D Euler (incompressible, periodic): the package `euler_v2_uniform`. Set EULER_DATA (the package directory:
# trajectories_fine.npz, split.json, normalization.npz) and EULER_TRUTH_BANK (group_<gid>.npz of the ten test
# conditions, 10,000 realizations each).
name: euler_v2_uniform
system: euler
data: ${EULER_DATA}
truth_bank: ${EULER_TRUTH_BANK}
expected_sha256: {split.json: 0040ecf525ba7c6b5cfe7d874812e04c69ccc4396ddbdc2db09e5489d8280b6e, normalization.npz: 06573da0f073b5bfabf884ac497f642c5d717824afd5644272239030dd7d4a7f,
trajectories_fine.npz: 664c3be66eee8b63dbad1fb379d72f6c7dd9a1094e3125c1e8c4223a4e2e7e66}
state:
dim: 196
observed: coarse spectral vorticity, |k| <= 8, Hermitian half (98 complex modes), omega_hat / N_sim^2
hidden: the unresolved vorticity 8 < |k| <= 64 of the initial condition
time_grid: {n_snapshots: 41, dt: 0.025, T: 1.0}
package: {vecs_train: [80, 200, 41, 196], vecs_eval: [20, 100, 41, 196], vecs_new: [720, 20, 41, 196]}
splits:
train: 80 conditions x 200 realizations (vecs_train); vecs_new holds 720 further conditions x 20 realizations, unused by the paper's models
val: 10 conditions x 100 realizations
test: 10 conditions x 100 realizations in the package; 10,000 realizations each in the truth bank
normalization: per-coordinate mean and std of the 80 training conditions x realizations j < 20 x all 41 snapshots
fid_adapter: {name: euler_coarse_vorticity, params: {n_disp: 128, k_c: 8}}