Convolutional neural network image classifier. Optimised to decipher between Darwin Nunez and Edinson Cavani.
%%%%% PERFORMANCE METRICS %%%%%%
TRAINING AND TEST SUMMARY:
Training Loss: 0.2644 Training Accuracy: 0.8865
Test Loss: 0.6664938926696777 Test Accuracy: 0.7250000238418579
ADDITIONAL TEST METRICS:
Accuracy: 0.725 Precision: 0.6956521739130435 Recall: 0.8 F1 Score: 0.7441860465116279
%%%%% PERFORMANCE METRICS %%%%%%
% % %
%%%%% ADDITIONAL DETAILS %%%%%%
DATA FORMAT:
Train Data Shape: (1851, 150, 150, 3) Train Labels Shape: (1851,) Test Data Shape: (200, 150, 150, 3) Test Labels Shape: (200,)
CNN MODEL STRUCTURE:
Model: "sequential"
conv2d (Conv2D) (None, 148, 148, 64) 1792
max_pooling2d (MaxPooling2D (None, 74, 74, 64) 0
)
conv2d_1 (Conv2D) (None, 72, 72, 128) 73856
max_pooling2d_1 (MaxPooling (None, 36, 36, 128) 0
2D)
conv2d_2 (Conv2D) (None, 34, 34, 256) 295168
max_pooling2d_2 (MaxPooling (None, 17, 17, 256) 0
2D)
conv2d_3 (Conv2D) (None, 15, 15, 512) 1180160
max_pooling2d_3 (MaxPooling (None, 7, 7, 512) 0
2D)
flatten (Flatten) (None, 25088) 0
dropout (Dropout) (None, 25088) 0
dense (Dense) (None, 512) 12845568
dense_1 (Dense) (None, 1) 513
================================================================= Total params: 14,397,057 Trainable params: 14,397,057 Non-trainable params: 0
TRAINING EPOCHS:
Epoch 1/10 47/47 [==============================] - 68s 1s/step - loss: 0.7232 - accuracy: 0.5946 - val_loss: 0.8840 - val_accuracy: 0.0000e+00 Epoch 2/10 47/47 [==============================] - 68s 1s/step - loss: 0.6721 - accuracy: 0.6061 - val_loss: 0.9562 - val_accuracy: 0.0054 Epoch 3/10 47/47 [==============================] - 66s 1s/step - loss: 0.6672 - accuracy: 0.6054 - val_loss: 0.6581 - val_accuracy: 0.6442 Epoch 4/10 47/47 [==============================] - 66s 1s/step - loss: 0.6647 - accuracy: 0.6250 - val_loss: 1.1434 - val_accuracy: 0.0027 Epoch 5/10 47/47 [==============================] - 66s 1s/step - loss: 0.5757 - accuracy: 0.7122 - val_loss: 0.8705 - val_accuracy: 0.5040 Epoch 6/10 47/47 [==============================] - 66s 1s/step - loss: 0.4689 - accuracy: 0.7784 - val_loss: 0.6468 - val_accuracy: 0.7035 Epoch 7/10 47/47 [==============================] - 66s 1s/step - loss: 0.4251 - accuracy: 0.8176 - val_loss: 0.9935 - val_accuracy: 0.5472 Epoch 8/10 47/47 [==============================] - 66s 1s/step - loss: 0.4077 - accuracy: 0.8250 - val_loss: 0.3853 - val_accuracy: 0.8032 Epoch 9/10 47/47 [==============================] - 66s 1s/step - loss: 0.3392 - accuracy: 0.8480 - val_loss: 0.1937 - val_accuracy: 0.9191 Epoch 10/10 47/47 [==============================] - 66s 1s/step - loss: 0.3381 - accuracy: 0.8486 - val_loss: 0.5162 - val_accuracy: 0.7574