Model Screening Resultaten

Dia 5/12

🏆 Model Screening Resultaten

4 CNN Architecturen Getest • 3 Epochs per Model • Vergelijking Performance

🥇 WINNER

EfficientNetB0

84.5%

Validation Accuracy

4.38M

Parameters

0.77 min

Training Time

1.097

Efficiency Score

238

Model Layers

EfficientNetB3

78.2%

Validation Accuracy

11.2M

Parameters

1.45 min

Training Time

0.539

Efficiency Score

385

Model Layers

ResNet50V2

76.6%

Validation Accuracy

24.1M

Parameters

1.33 min

Training Time

0.576

Efficiency Score

190

Model Layers

MobileNetV2

58.7%

Validation Accuracy

2.6M

Parameters

0.61 min

Training Time

0.962

Efficiency Score

154

Model Layers

🔍 Key Insights

⚡

Snelste

MobileNetV2
0.61 minuten

🎯

Beste Balans

EfficientNetB0
Accuracy + Efficiency

🏗️

Grootste

ResNet50V2
24.1M parameters