Adding one solution that produced a contact discontinuity

README.md

@@ -80,6 +80,10 @@ with polynomial refinement
 ```
 python examples/high_order_euler.py mlp.hidden.width=20 max_epochs=10000 mlp.segments=4 mlp.n=3 mlp.hidden.layers=8 factor=0.025 mlp.layer_type=continuous optimizer.patience=200 mlp.input.segments=20 batch_size=2048 form=primitive loss_weight.discontinuity=0.0 loss_weight.interior=1.0e-1 optimizer=adamw mlp.normalize=True mlp.rotations=6 gradient_clip=5.0 loss_weight.boundary=10 loss_weight.initial=10 data_size=10000 mlp.resnet=False refinement.type=p_refine refinement.epochs=1000
 ```
+this one actually got the contact discontinuity, but the velocity is wrong
+```
+python examples/high_order_euler.py mlp.hidden.width=10 max_epochs=10000 mlp.segments=2 mlp.n=3 mlp.hidden.layers=8 factor=0.025 mlp.layer_type=continuous optimizer.patience=200 mlp.input.segments=20 batch_size=2048 form=primitive loss_weight.discontinuity=0.0 loss_weight.interior=1.0e-1 optimizer=adamw mlp.normalize=True mlp.rotations=4 gradient_clip=5.0e-1 loss_weight.boundary=10 loss_weight.initial=10 data_size=10000 mlp.resnet=False refinement.type=p_refine refinement.epochs=1000
+```
 ## Training
 High order MLP
 ```

examples/high_order_euler.py

@@ -43,7 +43,11 @@ def run(cfg: DictConfig):
 
             # diff = cfg.mlp.target_n - cfg.mlp.n
             model = Net(cfg)
-
+            cfg.mlp.n = cfg.refinement.start_n
+            n = cfg.mlp.n
+            cfg.mlp.n_in = n
+            cfg.mlp.n_out = n
+            cfg.mlp.n_hidden = n
             for order in range(cfg.refinement.start_n, cfg.refinement.target_n):
                 trainer = Trainer(
                     max_epochs=cfg.refinement.epochs,
@@ -54,7 +58,12 @@ def run(cfg: DictConfig):
                 print(f"Training order {order}")
                 trainer.fit(model)
                 # trainer.test(model)
-                cfg.mlp.n = order + 1
+                n = order + 1
+                cfg.mlp.n = n
+                cfg.mlp.n_in = n
+                cfg.mlp.n_out = n
+                cfg.mlp.n_hidden = n
+
                 next_model = Net(cfg)
 
                 interpolate_high_order_mlp(network_in=model, network_out=next_model)