Test Results (20250221_173609)
==================================================

Train size:  1120
Val size:  160
Test size:  320
Trainset label distribution:  Counter({0: 119, 8: 117, 1: 115, 3: 113, 6: 110, 4: 110, 9: 109, 2: 109, 7: 109, 5: 109})
Valset label distribution:  Counter({2: 21, 7: 20, 4: 19, 8: 18, 5: 16, 0: 15, 1: 15, 9: 14, 6: 13, 3: 9})
Testset label distribution:  Counter({3: 38, 6: 37, 9: 37, 5: 35, 7: 31, 4: 31, 1: 30, 2: 30, 0: 26, 8: 25})
GAT initialized with 3 layers, [8, 8, 1] heads per layer, [34, 64, 64, 10] features per layer
Dropout: 0.1
Add skip connection: True
Bias: True
Log attention weights: True

Overall Accuracy: 0.9437

Per-class Metrics:
------------------------------

Class 0:
Precision: 0.9600
Recall: 0.9231
F1-score: 0.9412

Class 1:
Precision: 1.0000
Recall: 0.9667
F1-score: 0.9831

Class 2:
Precision: 1.0000
Recall: 0.8667
F1-score: 0.9286

Class 3:
Precision: 0.8605
Recall: 0.9737
F1-score: 0.9136

Class 4:
Precision: 0.9355
Recall: 0.9355
F1-score: 0.9355

Class 5:
Precision: 1.0000
Recall: 0.9429
F1-score: 0.9706

Class 6:
Precision: 0.8947
Recall: 0.9189
F1-score: 0.9067

Class 7:
Precision: 0.9394
Recall: 1.0000
F1-score: 0.9688

Class 8:
Precision: 0.9600
Recall: 0.9600
F1-score: 0.9600

Class 9:
Precision: 0.9459
Recall: 0.9459
F1-score: 0.9459
