def fit(self, training_data_file, training_label_file, num_classes, pixel):
train_set = read_data_sets(training_data_file, training_label_file,
num_classes=num_classes, one_hot=True)
self.session = tf.InteractiveSession()
# Create the model
self.x = tf.placeholder(tf.float32, [None, pixel])
self.W = tf.Variable(tf.zeros([pixel, num_classes]))
self.b = tf.Variable(tf.zeros([num_classes]))
self.y = tf.nn.softmax(tf.matmul(self.x, self.W) + self.b)
# Define loss and optimizer
y_ = tf.placeholder(tf.float32, [None, num_classes])
cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(self.y), reduction_indices=[1]))
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)
# Train
tf.initialize_all_variables().run()
for i in range(1000):
batch_xs, batch_ys = train_set.next_batch(200)
train_step.run({self.x: batch_xs, y_: batch_ys})
def png_to_features(filename):
width, height, pngdata, meta = png.Reader(filename).asDirect()
image = padding(13, 16, (width, height, numpy.vstack(map(numpy.uint8, pngdata))))
image = image.reshape((1, 13 * 16))
image = image.astype(numpy.float32)
image = numpy.multiply(image, 1.0 / 255.0)
return image
if __name__ == '__main__':
num_classes = 9
pixel = 208
train_set = read_data_sets('/home/zhou/train.gz',
'/home/zhou/train_label.gz', num_classes=num_classes, one_hot=True)
# test_set = read_data_sets('/home/zhou/test.gz')
sess = tf.InteractiveSession()
# Create the model
x = tf.placeholder(tf.float32, [None, pixel])
W = tf.Variable(tf.zeros([pixel, num_classes]))
b = tf.Variable(tf.zeros([num_classes]))
y = tf.nn.softmax(tf.matmul(x, W) + b)
# Define loss and optimizer
y_ = tf.placeholder(tf.float32, [None, num_classes])
cross_entropy = tf.reduce_mean(-tf.reduce_sum(y_ * tf.log(y), reduction_indices=[1]))
train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)
# Train
