# coding: utf-8
# ### Generating human faces with Adversarial Networks
import sys
sys.path.append("..")
import helpers
helpers.mask_busy_gpus(wait=False)
import numpy as np
#Those attributes will be required for the final part of the assignment (applying smiles), so please keep them in mind
#from lfw_dataset2 import load_lfw_dataset
from lfw_dataset import load_lfw_dataset
data, attrs = load_lfw_dataset(dimx=36, dimy=36)
#data = load_lfw_dataset(use_raw=True,dimx=36,dimy=36)
#print(np.max(data),np.min(data))
#preprocess faces
#data = np.float32(data)
#print(data[0])
data = (data - 127.5) / float(127.5) #scale to between -1 and 1
#print(data[0])
IMG_SHAPE = data.shape[1:]
# In[3]:
#print random image
print(data.shape)
import tensorflow as tf
gpu_options = tf.GPUOptions(allow_growth=True,
per_process_gpu_memory_fraction=0.333)
from keras_utils import reset_tf_session
# !!! remember to clear session/graph if you rebuild your graph to avoid out-of-memory errors !!!
"""# Load dataset
Dataset was downloaded for you. Relevant links (just in case):
- http://www.cs.columbia.edu/CAVE/databases/pubfig/download/lfw_attributes.txt
- http://vis-www.cs.umass.edu/lfw/lfw-deepfunneled.tgz
- http://vis-www.cs.umass.edu/lfw/lfw.tgz
"""
# we downloaded them for you, just link them here
download_utils.link_week_4_resources()
# load images
X, attr = load_lfw_dataset(use_raw=True, dimx=32, dimy=32)
IMG_SHAPE = X.shape[1:]
# center images
X = X.astype('float32') / 255.0 - 0.5
# split
X_train, X_test = train_test_split(X, test_size=0.1, random_state=42)
def show_image(x):
plt.imshow(np.clip(x + 0.5, 0, 1))
plt.title('sample images')
for i in range(6):
plt.subplot(2,3,i+1)
