from keras.models import load_model

# load model
model = load_model('model.h5')

# summarize model
print(model.summary())

# use model to predict something
predicted = model.predict(...)

# save model and architecture to single file
model.save('model.h5')
print('Saved model to disk')