model for detecting driver status

src/model.py

@@ -0,0 +1,50 @@
+import os
+import numpy as np
+import random, shutil
+
+from tensorflow.keras.preprocessing import image
+from tensorflow.keras.models import Sequential, load_model
+from tensorflow.keras.layers import Dropout, Conv2D, Flatten, Dense, MaxPool2D, BatchNormalization
+
+def generator(dir, gen=image.ImageDataGenerator(rescale=1./255), shuffle=True, batch_size=1, target_size=(24, 24), class_mode='categorical'):
+    return gen.flow_from_directory(dir, batch_size=batch_size, shuffle=shuffle, color_mode='grayscale', class_mode=class_mode, target_size=target_size)
+
+BS = 32
+TS = (24, 24)
+train_batch = generator('src/data/train', shuffle=True, batch_size=BS, target_size=TS)
+valid_batch = generator('src/data/valid', shuffle=True, batch_size=BS, target_size=TS)
+SPE = len(train_batch.classes)//BS
+VS = len(valid_batch.classes)//BS
+print(SPE,VS)
+
+model = Sequential([
+    Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=(24,24,1)),
+    MaxPool2D(pool_size=(1,1)),
+    Conv2D(32,(3,3),activation='relu'),
+    MaxPool2D(pool_size=(1,1)),
+#32 convolution filters used each of size 3x3
+#again
+    Conv2D(64, (3, 3), activation='relu'),
+    MaxPool2D(pool_size=(1,1)),
+
+#64 convolution filters used each of size 3x3
+#choose the best features via pooling
+
+#randomly turn neurons on and off to improve convergence
+    Dropout(0.25),
+#flatten since too many dimensions, we only want a classification output
+    Flatten(),
+#fully connected to get all relevant data
+    Dense(128, activation='relu'),
+#one more dropout for convergence' sake :) 
+    Dropout(0.5),
+#output a softmax to squash the matrix into output probabilities
+    Dense(2, activation='softmax')
+])
+
+model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])
+
+model.fit_generator(train_batch, validation_data=valid_batch,epochs=15, steps_per_epoch=SPE, validation_steps=VS)
+
+model.save('src/models/cnnCat2.h5', overwrite=True)
+