cifar10_checkpoint_improvements.py

from sklearn.preprocessing import LabelBinarizer
from pyimagesearch.nn.conv.minivggnet import MiniVGGNet
from keras.callbacks import ModelCheckpoint
from keras.optimizers import SGD
from keras.datasets import cifar10
import argparse
import os


ap = argparse.ArgumentParser()
ap.add_argument("-w", "--weights", required=True,
    help="path to weights directory")
args = vars(ap.parse_args())

print("[INFO] loading CIFAR-10 data...")

((trainX, trainY), (testX, testY)) = cifar10.load_data()

trainX = trainX.astype("float") / 255.0
testX = testX.astype("float") / 255.0

lb = LabelBinarizer()
trainY = lb.fit_transform(trainY)
testY = lb.transform(testY)

print("[INFO] compiling model")

opt = SGD(lr=0.01, decay=0.01/40, momentum=0.9, nesterov=True)
model = MiniVGGNet.build(width=32, height=32, depth=3, classes=10)
model.compile(loss="categorical_crossentropy", optimizer=opt,
    metrics=["accuracy"])

fname = os.path.sep.join([
    args["weights"],
    "weights-{epoch:03d}-{val_loss:.4f}.hdf5"
])

checkpoint = ModelCheckpoint(fname, monitor="val_loss", mode="min",
    save_best_only=True, verbose=1)
callbacks = [checkpoint]

# train the network
print("[INFO] training network...")

model.fit(trainX, trainY, validation_data=(testX, testY), batch_size=64,
    epochs=100, callbacks=callbacks, verbose=1)