npz files

Project.toml

@@ -34,6 +34,7 @@ LuxCUDA = "d0bbae9a-e099-4d5b-a835-1c6931763bda"
 LuxDeviceUtils = "34f89e08-e1d5-43b4-8944-0b49ac560553"
 MLUtils = "f1d291b0-491e-4a28-83b9-f70985020b54"
 NNlib = "872c559c-99b0-510c-b3b7-b6c96a88d5cd"
+NPZ = "15e1cf62-19b3-5cfa-8e77-841668bca605"
 NonlinearSolve = "8913a72c-1f9b-4ce2-8d82-65094dcecaec"
 Optimisers = "3bd65402-5787-11e9-1adc-39752487f4e2"
 Optimization = "7f7a1694-90dd-40f0-9382-eb1efda571ba"

examples/advect_fourier1D/advect1d_compare.jl

@@ -1,6 +1,7 @@
 #
 using GeometryLearning
 using Plots, LaTeXStrings
+using NPZ
 
 begin
     snfpath = joinpath(pkgdir(GeometryLearning), "examples", "autodecoder.jl")
@@ -98,6 +99,7 @@ modeldir_SNFL = joinpath(@__DIR__, "model_SNFL_l_$(ll)")
 #==================#
 # evolve
 #==================#
+
 case = 1
 
 modelfile_CINR = joinpath(modeldir_CINR, "model_08.jld2")
@@ -121,21 +123,46 @@ ud2, up2 = dropdims.((ud2, up2); dims = 1)
 ud3, up3 = dropdims.((ud3, up3); dims = 1)
 
 #==================#
-# figure
+# save data
 #==================#
+filename = joinpath(@__DIR__, "advect1d.npz")
+
+dict = Dict(
+    "xdata" => x1,
+    "tdata" => t1,
+    "udata" => ud1,
 
+    "xCROM" => x1,
+    "tCROM" => t1,
+    "uCROM" => up1,
+
+    "xSNFW" => x2,
+    "tSNFW" => t2,
+    "uSNFW" => up2,
+
+    "xSNFW" => x3,
+    "tSNFW" => t3,
+    "uSNFW" => up3,
+)
+
+npzwrite(filename, dict)
+
+#==================#
+# figure
+#==================#
 plt = plot(;
     xlabel = L"x",
     ylabel = L"u(x, t)",
 )
 
 Nt = length(t1)
-ii = [1, Nt]
+i1 = [1, Nt]
+i2 = Nt
 
-plot!(plt, x1, ud1[:, ii], w = 4, s = :solid     , c = :black, label = "Ground truth")
-plot!(plt, x1, up1[:, ii], w = 4, s = :dashdot   , c = :green, label = "C-ROM")
-plot!(plt, x2, up2[:, ii], w = 4, s = :dashdotdot, c = :red  , label = "Smooth-NFW (ours)")
-plot!(plt, x3, up3[:, ii], w = 4, s = :dashdotdot, c = :blue , label = "Smooth-NFL (ours)")
+plot!(plt, x1, ud1[:, i1], w = 4, s = :solid     , c = :black, label = "Ground truth")
+plot!(plt, x1, up1[:, i2], w = 4, s = :dashdot   , c = :green, label = "C-ROM")
+plot!(plt, x2, up2[:, i2], w = 4, s = :dashdotdot, c = :red  , label = "Smooth-NFW (ours)")
+plot!(plt, x3, up3[:, i2], w = 4, s = :dashdotdot, c = :blue , label = "Smooth-NFL (ours)")
 
 pltname = joinpath(@__DIR__, "compare_l_$(latent)")
 png(plt, pltname)

examples/advect_fourier1D/model_CINR_l_04/statistics_01.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 1.00086 || TEST LOSS: 1.00086
+Epoch [0 / 0]	TRAIN LOSS: 1.0066255 || TEST LOSS: 1.0066255
 #======================#
 TRAIN STATS
-R² score:             -284.4416
-MSE (mean SQR error): 1.00086
-RMSE (Root MSE):      1.0004299
-MAE (mean ABS error): 0.6974292
-maxAE (max ABS error) 3.5756156
-Lipschitz bound:      1.7745923e8
+R² score:             -82.881905
+MSE (mean SQR error): 1.0066255
+RMSE (Root MSE):      1.0033073
+MAE (mean ABS error): 0.69994503
+maxAE (max ABS error) 3.6405697
+Lipschitz bound:      2.4122685e8
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             -284.4416
-MSE (mean SQR error): 1.00086
-RMSE (Root MSE):      1.0004299
-MAE (mean ABS error): 0.6974292
-maxAE (max ABS error) 3.5756156
-Lipschitz bound:      1.7745923e8
+R² score:             -82.881905
+MSE (mean SQR error): 1.0066254
+RMSE (Root MSE):      1.0033072
+MAE (mean ABS error): 0.69994503
+maxAE (max ABS error) 3.6405697
+Lipschitz bound:      2.4122685e8
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_02.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 0.0017113755 || TEST LOSS: 0.0017113755
+Epoch [0 / 0]	TRAIN LOSS: 0.008209503 || TEST LOSS: 0.008209503
 #======================#
 TRAIN STATS
-R² score:             0.9982695
-MSE (mean SQR error): 0.0017113755
-RMSE (Root MSE):      0.041368775
-MAE (mean ABS error): 0.019237747
-maxAE (max ABS error) 1.436725
-Lipschitz bound:      1.4249618e8
+R² score:             0.9917643
+MSE (mean SQR error): 0.008209503
+RMSE (Root MSE):      0.09060631
+MAE (mean ABS error): 0.035630822
+maxAE (max ABS error) 3.707441
+Lipschitz bound:      9.286738e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.9982695
-MSE (mean SQR error): 0.0017113755
-RMSE (Root MSE):      0.041368775
-MAE (mean ABS error): 0.019237747
-maxAE (max ABS error) 1.436725
-Lipschitz bound:      1.4249618e8
+R² score:             0.9917643
+MSE (mean SQR error): 0.008209503
+RMSE (Root MSE):      0.09060631
+MAE (mean ABS error): 0.035630822
+maxAE (max ABS error) 3.707441
+Lipschitz bound:      9.286738e7
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_03.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 6.536157e-5 || TEST LOSS: 6.536157e-5
+Epoch [0 / 0]	TRAIN LOSS: 0.0034195452 || TEST LOSS: 0.0034195452
 #======================#
 TRAIN STATS
-R² score:             0.9999346
-MSE (mean SQR error): 6.536157e-5
-RMSE (Root MSE):      0.00808465
-MAE (mean ABS error): 0.00431201
-maxAE (max ABS error) 0.13435459
-Lipschitz bound:      1.3623653e8
+R² score:             0.99660486
+MSE (mean SQR error): 0.0034195452
+RMSE (Root MSE):      0.058476876
+MAE (mean ABS error): 0.019320237
+maxAE (max ABS error) 3.6904504
+Lipschitz bound:      8.2932376e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.9999346
-MSE (mean SQR error): 6.536158e-5
-RMSE (Root MSE):      0.008084651
-MAE (mean ABS error): 0.00431201
-maxAE (max ABS error) 0.13435459
-Lipschitz bound:      1.3623653e8
+R² score:             0.99660486
+MSE (mean SQR error): 0.0034195452
+RMSE (Root MSE):      0.058476876
+MAE (mean ABS error): 0.019320237
+maxAE (max ABS error) 3.6904504
+Lipschitz bound:      8.2932376e7
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_04.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 7.224428e-6 || TEST LOSS: 7.224428e-6
+Epoch [0 / 0]	TRAIN LOSS: 5.144835e-5 || TEST LOSS: 5.1448347e-5
 #======================#
 TRAIN STATS
-R² score:             0.9999928
-MSE (mean SQR error): 7.224428e-6
-RMSE (Root MSE):      0.0026878298
-MAE (mean ABS error): 0.001544364
-maxAE (max ABS error) 0.037864923
-Lipschitz bound:      1.2534767e8
+R² score:             0.9999486
+MSE (mean SQR error): 5.144835e-5
+RMSE (Root MSE):      0.0071727503
+MAE (mean ABS error): 0.0039398465
+maxAE (max ABS error) 0.09619093
+Lipschitz bound:      7.698274e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.9999928
-MSE (mean SQR error): 7.224428e-6
-RMSE (Root MSE):      0.0026878295
-MAE (mean ABS error): 0.0015443641
-maxAE (max ABS error) 0.037864923
-Lipschitz bound:      1.2534767e8
+R² score:             0.9999486
+MSE (mean SQR error): 5.1448347e-5
+RMSE (Root MSE):      0.0071727503
+MAE (mean ABS error): 0.0039398465
+maxAE (max ABS error) 0.09619093
+Lipschitz bound:      7.698274e7
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_05.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 4.9862592e-6 || TEST LOSS: 4.9862597e-6
+Epoch [0 / 0]	TRAIN LOSS: 2.5673378e-5 || TEST LOSS: 2.5673377e-5
 #======================#
 TRAIN STATS
-R² score:             0.99999493
-MSE (mean SQR error): 4.9862592e-6
-RMSE (Root MSE):      0.0022329933
-MAE (mean ABS error): 0.0013593892
-maxAE (max ABS error) 0.027574897
-Lipschitz bound:      1.2537754e8
+R² score:             0.9999743
+MSE (mean SQR error): 2.5673378e-5
+RMSE (Root MSE):      0.0050668903
+MAE (mean ABS error): 0.0029047912
+maxAE (max ABS error) 0.06899381
+Lipschitz bound:      7.669285e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.99999493
-MSE (mean SQR error): 4.9862592e-6
-RMSE (Root MSE):      0.0022329933
-MAE (mean ABS error): 0.0013593893
-maxAE (max ABS error) 0.027574897
-Lipschitz bound:      1.2537754e8
+R² score:             0.9999743
+MSE (mean SQR error): 2.5673378e-5
+RMSE (Root MSE):      0.0050668903
+MAE (mean ABS error): 0.002904791
+maxAE (max ABS error) 0.06899381
+Lipschitz bound:      7.669285e7
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_06.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 2.6238504e-6 || TEST LOSS: 2.6238502e-6
+Epoch [0 / 0]	TRAIN LOSS: 1.9715682e-5 || TEST LOSS: 1.9715682e-5
 #======================#
 TRAIN STATS
-R² score:             0.9999974
-MSE (mean SQR error): 2.6238504e-6
-RMSE (Root MSE):      0.0016198304
-MAE (mean ABS error): 0.0009697268
-maxAE (max ABS error) 0.021978855
-Lipschitz bound:      1.2353598e8
+R² score:             0.9999802
+MSE (mean SQR error): 1.9715682e-5
+RMSE (Root MSE):      0.0044402345
+MAE (mean ABS error): 0.0026330187
+maxAE (max ABS error) 0.05634752
+Lipschitz bound:      7.678265e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.9999974
-MSE (mean SQR error): 2.6238504e-6
-RMSE (Root MSE):      0.0016198304
-MAE (mean ABS error): 0.0009697268
-maxAE (max ABS error) 0.021978855
-Lipschitz bound:      1.2353598e8
+R² score:             0.9999802
+MSE (mean SQR error): 1.9715682e-5
+RMSE (Root MSE):      0.0044402345
+MAE (mean ABS error): 0.002633019
+maxAE (max ABS error) 0.05634752
+Lipschitz bound:      7.678265e7
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_07.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 9.255608e-7 || TEST LOSS: 9.255608e-7
+Epoch [0 / 0]	TRAIN LOSS: 7.349436e-6 || TEST LOSS: 7.349437e-6
 #======================#
 TRAIN STATS
-R² score:             0.99999905
-MSE (mean SQR error): 9.255608e-7
-RMSE (Root MSE):      0.00096206076
-MAE (mean ABS error): 0.0006050249
-maxAE (max ABS error) 0.008881092
-Lipschitz bound:      1.2250549e8
+R² score:             0.99999267
+MSE (mean SQR error): 7.349436e-6
+RMSE (Root MSE):      0.0027109843
+MAE (mean ABS error): 0.0016720209
+maxAE (max ABS error) 0.03496924
+Lipschitz bound:      7.612253e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.99999905
-MSE (mean SQR error): 9.255607e-7
-RMSE (Root MSE):      0.00096206076
-MAE (mean ABS error): 0.0006050249
-maxAE (max ABS error) 0.008881092
-Lipschitz bound:      1.2250549e8
+R² score:             0.99999267
+MSE (mean SQR error): 7.349436e-6
+RMSE (Root MSE):      0.0027109843
+MAE (mean ABS error): 0.0016720208
+maxAE (max ABS error) 0.03496924
+Lipschitz bound:      7.612253e7
 
 #======================#

examples/advect_fourier1D/model_CINR_l_04/statistics_08.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 6.401627e-7 || TEST LOSS: 6.401627e-7
+Epoch [0 / 0]	TRAIN LOSS: 3.7097816e-6 || TEST LOSS: 3.709782e-6
 #======================#
 TRAIN STATS
-R² score:             0.99999934
-MSE (mean SQR error): 6.401627e-7
-RMSE (Root MSE):      0.00080010167
-MAE (mean ABS error): 0.00052187586
-maxAE (max ABS error) 0.007401824
-Lipschitz bound:      1.2230506e8
+R² score:             0.9999963
+MSE (mean SQR error): 3.7097816e-6
+RMSE (Root MSE):      0.0019260794
+MAE (mean ABS error): 0.0012574346
+maxAE (max ABS error) 0.023277521
+Lipschitz bound:      7.532881e7
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             0.99999934
-MSE (mean SQR error): 6.401626e-7
-RMSE (Root MSE):      0.0008001016
-MAE (mean ABS error): 0.0005218759
-maxAE (max ABS error) 0.007401824
-Lipschitz bound:      1.2230506e8
+R² score:             0.9999963
+MSE (mean SQR error): 3.7097816e-6
+RMSE (Root MSE):      0.0019260794
+MAE (mean ABS error): 0.0012574347
+maxAE (max ABS error) 0.023277521
+Lipschitz bound:      7.532881e7
 
 #======================#

examples/advect_fourier1D/model_SNFL_l_04/statistics_01.txt

@@ -1,21 +1,21 @@
-Epoch [0 / 0]	TRAIN LOSS: 1.0041584 || TEST LOSS: 1.0041585
+Epoch [0 / 0]	TRAIN LOSS: 1.0042998 || TEST LOSS: 1.0042998
 #======================#
 TRAIN STATS
-R² score:             -269310.53
-MSE (mean SQR error): 0.99998814
-RMSE (Root MSE):      0.999994
-MAE (mean ABS error): 0.702276
-maxAE (max ABS error) 3.2403333
-Lipschitz bound:      130972.56
+R² score:             -1.473573e6
+MSE (mean SQR error): 0.9999838
+RMSE (Root MSE):      0.99999183
+MAE (mean ABS error): 0.7000161
+maxAE (max ABS error) 3.2416978
+Lipschitz bound:      131354.47
 
 #======================#
 #======================#
 TEST  STATS
-R² score:             -269310.5
-MSE (mean SQR error): 0.99998814
-RMSE (Root MSE):      0.999994
-MAE (mean ABS error): 0.70227605
-maxAE (max ABS error) 3.2403333
-Lipschitz bound:      130972.56
+R² score:             -1.473573e6
+MSE (mean SQR error): 0.99998367
+RMSE (Root MSE):      0.99999183
+MAE (mean ABS error): 0.7000161
+maxAE (max ABS error) 3.2416978
+Lipschitz bound:      131354.47
 
 #======================#