Regression on 1-D transient heat conduction system (val)
0.0001MSE (Data)DAE-HardNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| DAE-HardNetBest Epoch=1560, num_epochs=5000, model_depth=4, hidden_dim=32, lr=0.001, num_points=5000 out of 10000, pinn_reg_factor=1, hardnet_reg_factor=1, taylor_offset=0.1, taylor_order=2, eta=0.005, newton_step_length=1, max_newton_iter=5, noise_std=1, noise_mean=0, noise_scale=0.012025.12 | 0.0001 | 0 | 0 | |
| MLPBest Epoch=4240, num_epochs=5000, model_depth=4, hidden_dim=32, lr=0.001, num_points=5000 out of 10000, pinn_reg_factor=1, hardnet_reg_factor=1, taylor_offset=0.1, taylor_order=2, eta=0.005, newton_step_length=1, max_newton_iter=5, noise_std=1, noise_mean=0, noise_scale=0.012025.12 | 0.0001 | — | 0.0175 | |
| PINNBest Epoch=3650, num_epochs=5000, model_depth=4, hidden_dim=32, lr=0.001, num_points=5000 out of 10000, pinn_reg_factor=1, hardnet_reg_factor=1, taylor_offset=0.1, taylor_order=2, eta=0.005, newton_step_length=1, max_newton_iter=5, noise_std=1, noise_mean=0, noise_scale=0.012025.12 | 0.0021 | — | 0.0002 |