def main(data_path=args['data_path'], train_from=train_from):
train_gen, test_gen = create_data_generator(data_path)
valid_x, valid_y = create_valid_data()
if train_from == 'trained_weights':
model = load_model_from_trained_weights(imagedims=IMAGE_DIMS, nb_classes=len(train_gen.class_indices),
weights=args['weight_path'],
freeze_until=freeze_until)
elif train_from == 'trained_model':
model = load_model_from_trained_model()
else:
model = load_models(imagedims=IMAGE_DIMS, nb_classes=len(train_gen.class_indices))
print('[INFO] compiling model...')
model.compile(loss="categorical_crossentropy", optimizer=OPT, metrics=["accuracy"])
checkpoint = ModelCheckpoint(filepath=args['save_model'], monitor='val_loss', verbose=0,
save_best_only=True, save_weights_only=False,
mode='auto', period=1)
stop_early = EarlyStopping(monitor='val_loss', min_delta=.0, patience=40, verbose=0, mode='auto')
if lr_finder_from == 'large_range_search':
'''Exponential lr finder,
USE THIS FOR A LARGE RANGE SEARCH
Uncomment the validation_data flag to reduce speed but get a better idea of the learning rate
'''
lr_finder = LRFinder(NUM_SAMPLES, BS, minimum_lr=1e-3, maximum_lr=10.,
lr_scale='exp',
validation_data=(valid_x, valid_y), # use the validation data for losses
validation_sample_rate=5,
save_dir='weights/', verbose=True)
elif lr_finder_from == 'close_range_search':
'''LINEAR lr finder,
USE THIS FOR A CLOSE RANGE SEARCH
Uncomment the validation_data flag to reduce speed but get a better idea of the learning rate
'''
lr_finder = LRFinder(NUM_SAMPLES, BS, minimum_lr=1e-5, maximum_lr=1e-2,
lr_scale='exp',
validation_data=(valid_x, valid_y), # use the validation data for losses
validation_sample_rate=5,
save_dir='weights/', verbose=True)
callbacks = [checkpoint, stop_early, lr_finder]
H = model.fit_generator(train_gen,
validation_data=(valid_x, valid_y),
epochs=EPOCHS,
#steps_per_epoch=209,
callbacks=callbacks,
verbose=1
)
lr_finder.plot_schedule(clip_beginning=10, clip_endding=5)
print('Using real-time data augmentation.')
# This will do preprocessing and realtime data augmentation:
datagen = ImageDataGenerator(
featurewise_center=False, # set input mean to 0 over the dataset
samplewise_center=False, # set each sample mean to 0
featurewise_std_normalization=False, # divide inputs by std of the dataset
samplewise_std_normalization=False, # divide each input by its std
zca_whitening=False, # apply ZCA whitening
rotation_range=0, # randomly rotate images in the range (degrees, 0 to 180)
width_shift_range=0, # randomly shift images horizontally (fraction of total width)
height_shift_range=0, # randomly shift images vertically (fraction of total height)
horizontal_flip=True, # randomly flip images
vertical_flip=False) # randomly flip images
# Compute quantities required for featurewise normalization
# (std, mean, and principal components if ZCA whitening is applied).
datagen.fit(X_train)
# Fit the model on the batches generated by datagen.flow().
model.fit_generator(datagen.flow(X_train, Y_train, batch_size=batch_size, shuffle=True),
steps_per_epoch=X_train.shape[0] // batch_size,
validation_data=(X_test, Y_test),
epochs=nb_epoch, verbose=1,
callbacks=[lr_finder, model_checkpoint])
lr_finder.plot_schedule(clip_beginning=10, clip_endding=5)
scores = model.evaluate(X_test, Y_test, batch_size=batch_size)
for score, metric_name in zip(scores, model.metrics_names):
print("%s : %0.4f" % (metric_name, score))
featurewise_center=False, # set input mean to 0 over the dataset
samplewise_center=True, # set each sample mean to 0
featurewise_std_normalization=
False, # divide inputs by std of the dataset
samplewise_std_normalization=True, # divide each input by its std
zca_whitening=False, # apply ZCA whitening
# randomly rotate images in the range (degrees, 0 to 180)
rotation_range=90,
# randomly shift images horizontally (fraction of total width)
width_shift_range=0,
# randomly shift images vertically (fraction of total height)
height_shift_range=0,
horizontal_flip=True, # randomly flip images
vertical_flip=True) # randomly flip images
img_gen.fit(X_train)
VGG16_model.fit_generator(img_gen.flow(X_train,
Y_train,
batch_size=n_batch,
shuffle=True),
steps_per_epoch=X_train.shape[0] // n_batch,
validation_data=(X_test, Y_test),
epochs=n_epochs,
verbose=1,
callbacks=[lrf])
lrf.plot_schedule(clip_beginning=10, clip_endding=5)
scores = VGG16_model.evaluate(X_test, Y_test, batch_size=n_batch)
for score, metric_name in zip(scores, VGG16_model.metrics_names):
print("%s : %0.4f" % (metric_name, score))