Regression on PDE with multiple solution illustrative example 2 (train)
0.0007MSE (Data)DAE-HardNet
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DAE-HardNetBest Epoch=4820, num_epochs=5000, model_depth=4, hidden_dim=32, lr=0.001, num_points=2500 out of 2500, pinn_reg_factor=1, hardnet_reg_factor=1, taylor_offset=0.01, taylor_order=2, eta=0.05, newton_step_length=1, max_newton_iter=10, noise_std=1, noise_mean=0, noise_scale=0.012025.12 | 0.0007 | 0.0009 | 0 | |
| MLPBest Epoch=4990, num_epochs=5000, model_depth=4, hidden_dim=32, lr=0.001, num_points=2500 out of 2500, pinn_reg_factor=1, hardnet_reg_factor=1, taylor_offset=0.01, taylor_order=2, eta=0.05, newton_step_length=1, max_newton_iter=10, noise_std=1, noise_mean=0, noise_scale=0.012025.12 | 0.001 | — | 19 | |
| PINNBest Epoch=4390, num_epochs=5000, model_depth=4, hidden_dim=32, lr=0.001, num_points=2500 out of 2500, pinn_reg_factor=1, hardnet_reg_factor=1, taylor_offset=0.01, taylor_order=2, eta=0.05, newton_step_length=1, max_newton_iter=10, noise_std=1, noise_mean=0, noise_scale=0.012025.12 | 0.0431 | — | 0.344 |