🎯 MNIST Classification Report

Generated on 2025-06-14 10:42:19

📊 Overall Performance

0.990
Test Accuracy
0.990
Avg Precision
0.990
Avg Recall
0.990
Avg F1-Score
10,000
Test Samples
0.992
Avg Confidence

🔢 Per-Digit Performance

Digit 0 0.994
Precision: 0.994 Recall: 0.994 F1: 0.994 Support: 980.0
Digit 1 0.994
Precision: 0.995 Recall: 0.994 F1: 0.994 Support: 1135.0
Digit 2 0.992
Precision: 0.993 Recall: 0.992 F1: 0.993 Support: 1032.0
Digit 3 0.998
Precision: 0.973 Recall: 0.998 F1: 0.985 Support: 1010.0
Digit 4 0.985
Precision: 0.997 Recall: 0.985 F1: 0.991 Support: 982.0
Digit 5 0.981
Precision: 0.993 Recall: 0.981 F1: 0.987 Support: 892.0
Digit 6 0.989
Precision: 0.990 Recall: 0.989 F1: 0.989 Support: 958.0
Digit 7 0.991
Precision: 0.987 Recall: 0.991 F1: 0.989 Support: 1028.0
Digit 8 0.984
Precision: 0.998 Recall: 0.984 F1: 0.991 Support: 974.0
Digit 9 0.990
Precision: 0.980 Recall: 0.990 F1: 0.985 Support: 1009.0

🎯 Confusion Matrix

Darker colors indicate higher values. Green diagonal shows correct predictions, red off-diagonal shows misclassifications.

Pred 0Pred 1Pred 2Pred 3Pred 4Pred 5Pred 6Pred 7Pred 8Pred 9
True 0974010004100
True 10112813010200
True 21010242000500
True 30001008010100
True 40010967011012
True 50001108754101
True 6331012947010
True 70122000101913
True 8221600109584
True 9000422020999

🏋️ Training Information

Training history not available.

Model Details

ArchitectureSimpleCNN
Parameters130,890
Epochs Trained5
Batch Size64
Learning Rate0.001
Devicecuda
Final Train Acc98.62%
Final Val Acc98.78%

💡 Analysis

🎉 Excellent Performance! The model achieved 99.0% accuracy on MNIST test set.

  • Best recognized digit: 3 (99.8% accuracy)
  • Most challenging digit: 5 (98.1% accuracy)
  • Model confidence: Average confidence of 99.2%
  • Training efficiency: Achieved 98.8% validation accuracy in 5 epochs

Report generated by ML Pipeline Orchestrator - MNIST Classification Pipeline