def print_best_model():
space_best_model = load_best_hyperspace()
if space_best_model is None:
print("No best model to plot. Continuing...")
return
print("Best hyperspace yet:")
print_model_data(['model_name','train_accuracy','test_accuracy','model_param_num'])
print_json(space_best_model)
def plot_best_model():
"""Plot the best model found yet."""
space_best_model = load_best_hyperspace()
if space_best_model is None:
tf.logging.info("No best model to plot. Continuing...")
return
tf.logging.info("Best hyperspace yet:")
print_json(space_best_model)
plot(space_best_model, "model_best")
def plot_best_model():
"""Plot the best model found yet."""
space_best_model = load_best_hyperspace()
if space_best_model is None:
print("No best model to plot. Continuing...")
return
# Print best hyperspace and save model png
print("Best hyperspace yet:")
print_json(space_best_model)
plot(space_best_model, "model_best")
import numpy as np
from keras import backend as K
import time
import os
# Dimensions of the generated pictures for each filter.
img_width = 32
img_height = 32
weight_file = "{}/f37d5.hdf5".format(WEIGHTS_DIR)
LAYERS_DIR = "layers"
# Load model in test phase mode: no dropout, and use fixed BN
K.set_learning_phase(0)
model = build_model(load_best_hyperspace())
model.load_weights(weight_file)
print('Model loaded.')
model.summary()
def normalize(x):
"""Utility function to normalize a tensor by its L2 norm."""
return x / (K.sqrt(K.mean(K.square(x))) + 1e-5)
def deprocess_image(x):
"""Utility function to convert a tensor into a valid image."""
# Normalize tensor: center on 0., ensure std is 0.1
x -= x.mean()
from neural_net import build_and_train, TENSORBOARD_DIR
from utils import print_json, load_best_hyperspace
from keras.layers.core import K
import os
__author__ = "Vooban Inc."
__copyright__ = "Copyright 2017, Vooban Inc."
__license__ = "MIT License"
# See: https://github.com/Vooban/Hyperopt-Keras-CNN-CIFAR-100/blob/master/LICENSE"
if __name__ == "__main__":
"""Retrain best model with TensorBoard. Also save best weights."""
space_best_model = load_best_hyperspace()
print("Hyperspace:")
print_json(space_best_model)
model, model_name, results, log_path = build_and_train(
space_best_model, save_best_weights=True, log_for_tensorboard=True)
print("Model Name:", model_name)
print("Note: results 'json' file not saved to 'results/' since it is now "
"available in TensorBoard. See above console output for json-styled "
"results.")
print("Model summary:")
model.summary()