def fetch_data(source, targets, amount, numthreads=10, threadtimeout=1):
faces = []
fclass = []
classes = to_one_hot(targets)
for k,target in enumerate(targets):
tfaces = sorted(set(glob("cropped/" + target + "/*")))
if len(tfaces) < amount or amount == 0:
fetch_data_files(source, [target], amount - len(tfaces), numthreads, threadtimeout)
tfaces = sorted(set(glob("cropped/" + target + "/*")))
for i in range(len(tfaces)):
face = imread(tfaces[i], mode='RGB')
faces.append(face)
fclass.append(classes[k])
return faces, fclass
def fetch_data(source, targets, amount, numthreads=10, threadtimeout=1):
faces = []
fclass = []
classes = to_one_hot(targets)
for k, target in enumerate(targets):
tfaces = sorted(set(glob("cropped/" + target + "/*")))
if len(tfaces) < amount or amount == 0:
fetch_data_files(source, [target], amount - len(tfaces),
numthreads, threadtimeout)
tfaces = sorted(set(glob("cropped/" + target + "/*")))
for i in range(len(tfaces)):
face = imread(tfaces[i], mode='RGB')
faces.append(face)
fclass.append(classes[k])
return faces, fclass
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
from random import shuffle
from set_utils import make_sets, to_one_hot
# Load data
with np.load("notMNIST.npz") as data:
images, labelso = data["images"], data["labels"]
images = images.transpose(2,0,1)
images = images.reshape(18720, 784)
poissonNoise = np.random.poisson(50,784).astype(float)
images = images.astype('float32')/255
labels = to_one_hot(labelso)
''' PARAMETERS '''
learning_rate = 1e-2
training_epochs = 600
batch_size = 500
momentum = 1e-2
hidden_units = 1000
''' SETS '''
x_train, t_train, x_validation, t_validation, x_test, t_test = make_sets(images, labels, 15000, 1000)
#NN Model
x = tf.placeholder(tf.float32, [None, 784])
y = tf.placeholder(tf.float32, [None, 10])
#Input Layer
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
from random import shuffle
from set_utils import make_sets, to_one_hot
# Load data
with np.load("notMNIST.npz") as data:
images, labelso = data["images"], data["labels"]
images = images.transpose(2, 0, 1)
images = images.reshape(18720, 784)
poissonNoise = np.random.poisson(50, 784).astype(float)
images = images.astype('float32') / 255
labels = to_one_hot(labelso)
''' PARAMETERS '''
learning_rate = 1e-2
training_epochs = 600
batch_size = 500
momentum = 1e-2
hidden_units = 1000
''' SETS '''
x_train, t_train, x_validation, t_validation, x_test, t_test = make_sets(
images, labels, 15000, 1000)
#NN Model
x = tf.placeholder(tf.float32, [None, 784])
y = tf.placeholder(tf.float32, [None, 10])
#Input Layer
input_w = tf.Variable(tf.truncated_normal([784, hidden_units], stddev=0.01),
name="Input_Weight")