folivo/PINA
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

modify tutorials for plotter compatibility

Browse Source
This commit is contained in:
Dario Coscia
2023-10-31 12:03:04 +01:00
committed by Nicola Demo
parent 5336f36f08
commit d556c592e0
36 changed files with 290 additions and 260 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
tutorials/tutorial2/tutorial.py vendored
View File

@@ -106,7 +106,7 @@ trainer.train()
plotter = Plotter()
plotter.plot(trainer)
plotter.plot(solver=pinn)
# ## Solving the problem with extra-features PINNs
@@ -160,7 +160,7 @@ trainer_feat.train()
# In[6]:
plotter.plot(trainer_feat)
plotter.plot(solver=pinn_feat)
# ## Solving the problem with learnable extra-features PINNs
@@ -177,7 +177,7 @@ plotter.plot(trainer_feat)
# where $\alpha$ and $\beta$ are the abovementioned parameters.
# Their implementation is quite trivial: by using the class `torch.nn.Parameter` we cam define all the learnable parameters we need, and they are managed by `autograd` module!
# In[8]:
# In[14]:
class SinSinAB(torch.nn.Module):
@@ -212,7 +212,7 @@ trainer_learn.train()
# Umh, the final loss is not appreciabily better than previous model (with static extra features), despite the usage of learnable parameters. This is mainly due to the over-parametrization of the network: there are many parameter to optimize during the training, and the model in unable to understand automatically that only the parameters of the extra feature (and not the weights/bias of the FFN) should be tuned in order to fit our problem. A longer training can be helpful, but in this case the faster way to reach machine precision for solving the Poisson problem is removing all the hidden layers in the `FeedForward`, keeping only the $\alpha$ and $\beta$ parameters of the extra feature.
# In[11]:
# In[19]:
# make model + solver + trainer
@@ -222,7 +222,7 @@ model_lean= FeedForward(
output_dimensions=len(problem.output_variables),
input_dimensions=len(problem.input_variables)+1
)
pinn_learn = PINN(problem, model_lean, extra_features=[SinSinAB()], optimizer_kwargs={'lr':0.006, 'weight_decay':1e-8})
pinn_learn = PINN(problem, model_lean, extra_features=[SinSinAB()], optimizer_kwargs={'lr':0.01, 'weight_decay':1e-8})
trainer_learn = Trainer(pinn_learn, max_epochs=1000, callbacks=[MetricTracker()], accelerator='cpu', enable_model_summary=False) # we train on CPU and avoid model summary at beginning of training (optional)
# train
@@ -234,20 +234,20 @@ trainer_learn.train()
#
# We conclude here by showing the graphical comparison of the unknown field and the loss trend for all the test cases presented here: the standard PINN, PINN with extra features, and PINN with learnable extra features.
# In[12]:
# In[20]:
plotter.plot(trainer_learn)
plotter.plot(solver=pinn_learn)
# Let us compare the training losses for the various types of training
# In[14]:
# In[21]:
plotter.plot_loss(trainer, label='Standard')
plotter.plot_loss(trainer_feat, label='Static Features')
plotter.plot_loss(trainer_learn, label='Learnable Features')
plotter.plot_loss(trainer, logy=True, label='Standard')
plotter.plot_loss(trainer_feat, logy=True,label='Static Features')
plotter.plot_loss(trainer_learn, logy=True, label='Learnable Features')
# ## What's next?