# Initialize the training and validation dataset generators
trainGen = HDF5DatasetGenerator(dbPath=data_config.TRAIN_HDF5,
batchSize=128,
preprocessor=[pp, mp, iap],
aug=None,
classes=2)
valGen = HDF5DatasetGenerator(dbPath=data_config.VAL_HDF5,
batchSize=128,
preprocessor=[sp, mp, iap],
classes=2)
# Initialize the optimizer
print("[INFO] compiling model..,")
opt = Adam(lr=1e-3)
model = AlexNet.build(width=227, height=227, depth=3, classes=2, reg=0.0002)
model.compile(loss="binary_crossentropy", optimizer=opt, metrics=["accuracy"])
# Construct the set of callbacks
path = os.path.sep.join(
[data_config.OUTPUT_PATH, "{}.png".format(os.getpid())])
callbacks = [TrainingMonitor(path)]
# train the network
model.fit_generator(trainGen.generator(),
steps_per_epoch=trainGen.numImages // 128,
validation_data=valGen.generator(),
validation_steps=valGen.numImages // 128,
epochs=75,
max_queue_size=128 * 2,
callbacks=None,
batch_size=batch_size,
class_mode='categorical',
subset='validation',
classes=CLASSES)
tbCallBack = TensorBoard(log_dir='./logs',
histogram_freq=0,
write_graph=True,
write_images=True)
# initialize the optimizer
print("[INFO] compiling model...")
opt = Adam(lr=0.5e-3)
model = AlexNet.build(width=height,
height=width,
depth=channel,
classes=num_class,
reg=0.0002)
model.compile(loss="categorical_crossentropy",
optimizer=opt,
metrics=["accuracy"])
# train the network
model.fit_generator(
train_generator,
steps_per_epoch=train_generator.samples // batch_size,
validation_data=validation_generator,
validation_steps=validation_generator.samples // batch_size,
epochs=nb_epochs,
# 0 = silent, 1 = progress bar, 2 = one line per epoch.
callbacks=[tbCallBack],
parser.add_argument("-nn", "--network", required=True, \
help="name of neural network")
parser.add_argument("-o", "--output", required=True, \
help="path to output pictures")
args = vars(parser.parse_args())
networkBanks = {
"vgg16": VGG16(weights="imagenet"),
"vgg19": VGG19(weights="imagenet"),
"kerasresnet50": KerasResNet50(weights="imagenet"),
"inceptionv3": InceptionV3(weights="imagenet"),
"shallownet": ShallowNet.build(height=28, width=28, depth=3, classes=10),
"lenet": LeNet.build(height=28, width=28, depth=3, classes=10),
"minivgg": MiniVGGNet.build(height=28, width=28, depth=3, classes=10),
#"kfer_lenet" : KFER_LeNet.build(height=48, width=48, depth=3, classes=7),
"alexnet": AlexNet.build(height=227, width=227, depth=3, classes=1000),
"alexnet2": AlexNet2.build(height=227, width=227, depth=3, classes=1000),
"resnet50": ResNet50(height=224, width=224, depth=3, classes=10),
"resnet101": ResNet101(height=224, width=224, depth=3, classes=10),
"resnet152": ResNet152(height=224, width=224, depth=3, classes=10),
}
model = networkBanks[args["network"]]
arch_path = os.path.join(args["output"], \
"{}_architecture.png".format(args["network"]))
"""
show_shapes = print out each layer's shape;
or use tf.compat.v1.keras.utils.plot_model() / tf.keras.utils.load_model(), with
rankdir = TB/LR, plotting layers in V/H direction
