updating the pre-commit

Author: bruAristimunha

examples/howto/plot_howto_add_batchnorm.py

@@ -25,6 +25,7 @@
 
 from spd_learn.modules import BiMap, LogEig, ReEig, SPDBatchNormLie
 
+
 ######################################################################
 # Step 1: Choose Your Normalization Layer
 # ----------------------------------------

examples/howto/plot_howto_choose_metric.py

@@ -45,8 +45,10 @@
 import torch
 
 from pyriemann.datasets import make_gaussian_blobs
+
 from spd_learn.modules import SPDBatchNormLie
 
+
 torch.manual_seed(42)
 
 # Generate SPD data using pyriemann (2-class, 2*n_matrices total samples)
@@ -87,7 +89,9 @@
         _ = bn(X_bench)
     elapsed = (time.time() - t0) / 20
     timings[metric] = elapsed * 1000
-    print(f"{metric}: {elapsed*1000:.1f} ms/batch ({n_bench}x{n_bench}, batch={batch_size})")
+    print(
+        f"{metric}: {elapsed * 1000:.1f} ms/batch ({n_bench}x{n_bench}, batch={batch_size})"
+    )
 
 fig, ax = plt.subplots(figsize=(6, 4))
 ax.bar(
@@ -177,7 +181,9 @@
         color=metric_colors[metric],
         alpha=0.8,
     )
-    ax.set_title(f"{metric} Eigenvalues", fontweight="bold", color=metric_colors[metric])
+    ax.set_title(
+        f"{metric} Eigenvalues", fontweight="bold", color=metric_colors[metric]
+    )
     ax.set_xlabel("Index")
 
 plt.suptitle(
@@ -231,8 +237,7 @@
     ax.grid(True, alpha=0.3)
 
     print(
-        f"AIM (theta={theta}): eigval range "
-        f"[{eigvals.min():.3f}, {eigvals.max():.3f}]"
+        f"AIM (theta={theta}): eigval range [{eigvals.min():.3f}, {eigvals.max():.3f}]"
     )
 
 axes[0].set_ylabel("Eigenvalue")

examples/tutorials/tutorial_05_batch_normalization.py

@@ -39,6 +39,7 @@
     SPDBatchNormMeanVar,
 )
 
+
 torch.manual_seed(42)
 np.random.seed(42)
 
@@ -224,8 +225,12 @@ def train_model(model, X_train, y_train, X_test, y_test, epochs=150, lr=5e-3):
 for metric, (_, accs) in liebn_results.items():
     epochs, vals = zip(*accs)
     ax2.plot(
-        epochs, vals, "o-", label=f"LieBN ({metric})",
-        color=colors[metric], markersize=3,
+        epochs,
+        vals,
+        "o-",
+        label=f"LieBN ({metric})",
+        color=colors[metric],
+        markersize=3,
     )
 ax2.set_xlabel("Epoch")
 ax2.set_ylabel("Test Accuracy")
@@ -328,8 +333,7 @@ def train_model(model, X_train, y_train, X_test, y_test, epochs=150, lr=5e-3):
     loss.backward()
     eigvals = torch.linalg.eigvalsh(out.detach())
     print(
-        f"{metric}: min_eigval={eigvals.min():.2e}, "
-        f"grad_norm={X_check.grad.norm():.4f}"
+        f"{metric}: min_eigval={eigvals.min():.2e}, grad_norm={X_check.grad.norm():.4f}"
     )
     X_check.grad = None