edgarmoyete / cifar

Clasificacion de imagenes CIFAR10

Clasificacion de imagenes con el conjunto de datos CIFAR10

CNN.ipynb

• lr = por defecto
• epochs = 20
• accuracy = 72.83%

CNN_BN.ipynb

• lr = por defecto
• epochs = 20
• accuracy = 72.17%

DataAugmentation.ipynb

• lr = por defecto
• epochs = 100
• accuracy = 77.27%

Transfer_Learning.ipynb

• lr = 1e-4
• epochs = 200
• accuracy = 61.66%

Transfer_Learning_2.ipynb

• lr = 1e-4
• epochs = 200
• accuracy = 61.03%

Fine_Tuning.ipynb

• lr = 1e-6
• epochs = 200
• accuracy = 75.91%

Fine_Tuning_DataAugmentation.ipynb

• lr = 1e-6
• epochs = 200
• accuracy = 72.54%

Fine_Tuning_2.ipynb

• lr = 1e-4, 1e-6
• epochs = 100, 100
• accuracy = 72.86%

FineTuning2_DataAugmentation.ipynb

• lr = 1e-4, 1e-6
• epochs = 100, 100
• accuracy = 69.63%

DataAugmentationExplorar.ipynb

• Ejemplo de aumento de datos