folivo/thermal-conduction-ml
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

new data format

Browse Source
This commit is contained in:
Filippo Olivo
2025-12-12 10:18:16 +01:00
parent 27c2aeb736
commit 732d48c360
3 changed files with 118 additions and 87 deletions
Download Patch File Download Diff File Expand all files Collapse all files

128
ThermalSolver/autoregressive_module.py
View File

@@ -16,48 +16,79 @@ def import_class(class_path: str):
def _plot_mesh(pos_, y_, y_pred_, y_true_, batch, i, batch_idx):
for j in [0, 10, 20, 30]:
# print(pos_.shape, y_.shape, y_pred_.shape, y_true_.shape)
for j in [0]:
idx = (batch == j).nonzero(as_tuple=True)[0]
y = y_[idx].detach().cpu()
y_pred = y_pred_[idx].detach().cpu()
pos = pos_[idx].detach().cpu()
# print(pos.shape, y.shape, y_pred.shape)
y_true = y_true_[idx].detach().cpu()
y_true = torch.clamp(y_true, min=0)
folder = f"{j:02d}_images"
if os.path.exists(folder) is False:
os.makedirs(folder)
pos = pos.detach().cpu()
tria = Triangulation(pos[:, 0], pos[:, 1])
plt.figure(figsize=(24, 5))
plt.subplot(1, 4, 1)
plt.tricontourf(tria, y.squeeze().numpy(), levels=100)
plt.colorbar()
plt.title("Step t-1")
plt.subplot(1, 4, 2)
plt.tricontourf(tria, y_pred.squeeze().numpy(), levels=100)
plt.figure(figsize=(18, 6))
# plt.subplot(1, 4, 1)
# plt.tricontourf(tria, y.squeeze().numpy(), levels=100)
# plt.colorbar()
# plt.title("Step t-1")
# plt.tripcolor(tria, y_pred.squeeze().numpy()
# plt.savefig("test_scatter_step_before.png", dpi=72)
# x = z
plt.subplot(1, 3, 1)
# plt.tricontourf(tria, y_pred.squeeze().numpy(), levels=100)
plt.scatter(
pos[:, 0],
pos[:, 1],
c=y_pred.squeeze().numpy(),
s=20,
cmap="viridis",
)
plt.colorbar()
plt.title("Step t Predicted")
plt.subplot(1, 4, 3)
plt.tricontourf(tria, y_true.squeeze().numpy(), levels=100)
plt.subplot(1, 3, 2)
# plt.tricontourf(tria, y_true.squeeze().numpy(), levels=100)
plt.scatter(
pos[:, 0],
pos[:, 1],
c=y_true.squeeze().numpy(),
s=20,
cmap="viridis",
)
plt.colorbar()
plt.title("t True")
plt.subplot(1, 4, 4)
plt.tricontourf(tria, (y_true - y_pred).squeeze().numpy(), levels=100)
plt.subplot(1, 3, 3)
per_element_relative_error = torch.abs(y_pred - y_true) / torch.clamp(
torch.abs(y_true), min=1e-6
)
# plt.tricontourf(tria, per_element_relative_error.squeeze(), levels=100)
plt.scatter(
pos[:, 0],
pos[:, 1],
c=per_element_relative_error.squeeze().numpy(),
s=20,
cmap="viridis",
)
plt.colorbar()
plt.title("Error")
plt.title("Relative Error")
plt.suptitle("GNO", fontsize=16)
name = f"{folder}/{j:04d}_graph_iter_{i:04d}.png"
plt.savefig(name, dpi=72)
plt.close()
def _plot_losses(losses, batch_idx):
def _plot_losses(test_losses, batch_idx):
folder = f"{batch_idx:02d}_images"
plt.figure()
plt.plot(losses)
for i, losses in enumerate(test_losses):
plt.plot(losses)
if i == 3:
break
plt.yscale("log")
plt.xlabel("Iteration")
plt.ylabel("Loss")
plt.ylabel("Relative Error")
plt.title("Test Loss over Iterations")
plt.grid(True)
file_name = f"{folder}/test_loss.png"
@@ -80,6 +111,7 @@ class GraphSolver(LightningModule):
# print(f"Param: {param[0]}")
self.loss = loss if loss is not None else torch.nn.MSELoss()
self.unrolling_steps = unrolling_steps
self.test_losses = []
def _compute_loss(self, x, y):
return self.loss(x, y)
@@ -149,7 +181,7 @@ class GraphSolver(LightningModule):
self._log_loss(loss, batch, "train")
for i, layer in enumerate(self.model.layers):
self.log(
f"alpha_{i}",
f"{i:03d}_alpha",
layer.alpha,
prog_bar=True,
on_epoch=True,
@@ -205,10 +237,10 @@ class GraphSolver(LightningModule):
self._log_loss(loss, batch, "val")
return loss
def _check_convergence(self, y_pred, y_true, tol=1e-3):
l2_norm = torch.norm(y_pred - y_true, p=2)
y_true_norm = torch.norm(y_true, p=2)
rel_error = l2_norm / (y_true_norm + 1e-8)
def _check_convergence(self, y_new, y_old, tol=1e-3):
l2_norm = torch.norm(y_new, p=2) - torch.norm(y_old, p=2)
y_old_norm = torch.norm(y_old, p=2)
rel_error = l2_norm / (y_old_norm)
return rel_error.item() < tol
def test_step(self, batch: Batch, batch_idx):
@@ -219,7 +251,9 @@ class GraphSolver(LightningModule):
losses = []
all_losses = []
norms = []
for i in range(self.unrolling_steps):
sequence_length = y.size(1)
y = y[:, -1, :].unsqueeze(1)
for i in range(100):
out = self._compute_model_steps(
# torch.cat([x,pos], dim=-1),
x,
@@ -231,34 +265,38 @@ class GraphSolver(LightningModule):
conductivity,
)
norms.append(torch.norm(out - x, p=2).item())
converged = self._check_convergence(out, x)
if batch_idx == 0:
_plot_mesh(
batch.pos,
x,
out,
y[:, -1, :],
batch.batch,
i,
self.current_epoch,
)
x = out
loss = self.loss(out, y[:, i, :])
all_losses.append(loss.item())
loss = self.loss(out, y[:, -1, :])
relative_error = torch.norm(out - y[:, -1, :], p=2) / torch.norm(
y[:, -1, :], p=2
)
all_losses.append(relative_error.item())
losses.append(loss)
# if (
# batch_idx == 0
# and self.current_epoch % 10 == 0
# and self.current_epoch > 0
# ):
# _plot_mesh(
# batch.pos,
# x,
# out,
# y[:, i, :],
# batch.batch,
# i,
# self.current_epoch,
# )
if converged:
print(
f"Test step converged at iteration {i} for batch {batch_idx}"
)
break
loss = torch.stack(losses).mean()
# if (
# batch_idx == 0
# and self.current_epoch % 10 == 0
# and self.current_epoch > 0
# ):
_plot_losses(norms, self.current_epoch)
self.test_losses.append(all_losses)
self._log_loss(loss, batch, "test")
return loss
def on_test_end(self):
if len(self.test_losses) > 0:
_plot_losses(self.test_losses, batch_idx=0)
def configure_optimizers(self):
optimizer = torch.optim.AdamW(self.parameters(), lr=1e-3)
return optimizer