DataScience Lab 2017_Мониторинг модных трендов с помощью глубокого обучения и TensorFlow_Ольга Романюк
1 like•233 views
The document discusses deep residual learning techniques for image recognition as proposed by Kaiming He et al. It emphasizes the construction of residual networks, the use of identity mappings, and optimizes training through techniques such as batch normalization and adaptive learning rates. Practical implementations using TensorFlow are outlined, detailing the architecture and training process for classification tasks on datasets like MNIST.
![graph = tf.Graph()
with graph.as_default():
x = tf.placeholder(tf.float32,name='input')
y = tf.placeholder(tf.float32,name='labels')
x_image = tf.reshape(x, [-1, 28, 28, 1])
...](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-9-320.jpg)
![with tf.variable_scope('conv1'):
W_conv1 = weight_variable(
[5, 5, 1, 32])
b_conv1 = bias_variable([32])
h_conv1 = tf.nn.relu(
conv2d(x_image, W_conv1) + b_conv1)
h_pool1 = max_pool_2x2(h_conv1)
with tf.variable_scope('conv2'):
...](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-10-320.jpg)
![with tf.Session(graph=graph) as sess:
from tf.examples.tutorials.mnist import input_data
mnist = input_data.read_data_sets('MNIST_data', one_hot=True)
writer = tf.summary.FileWriter(logdir, sess.graph)
sess.run(tf.global_variables_initializer())
for i in range(20000):
batch = mnist.train.next_batch(50)
_,train_accuracy = sess.run([train_step,accuracy],
feed_dict={x: batch[0], y: batch[1]})
if i % 100 == 0:
print("step %d, training accuracy %g" % (i, train_accuracy))
print("test accuracy %g" % sess.run(accuracy, feed_dict={x:
mnist.test.images, y: mnist.test.labels}))](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-12-320.jpg)
![def conv_layer_resnet_im(inpt, filter_shape, stride,
phase,name=''):
filter_ = weight_variable(filter_shape,
name=name + '_weights')
normalized = batch_norm(inpt, phase,name=name)
activated = tf.nn.relu(normalized)
conv = tf.nn.conv2d(activated, filter=filter_,
strides=[1, stride, stride, 1], padding="SAME")
return conv](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-14-320.jpg)
![def residual_block_im(inpt, output_depth,phase,
name=''):
input_depth = inpt.get_shape().as_list()[3]
conv1 = conv_layer_resnet_im(inpt,
[3, 3,input_depth, output_depth],
1,phase, name=name+'_conv1')
conv2 = conv_layer_resnet_im(conv1,
[3, 3, output_depth, output_depth],
1,phase, name=name+'_conv2')
if input_depth != output_depth:
input_layer = tf.pad(inpt, [[0,0], [0,0],
[0,0], [0, output_depth - input_depth]])
else:
input_layer = inpt
return conv2 + input_layer](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-15-320.jpg)
![num_blocks = 4
num_filters = 128
for i in range(num_blocks):
with tf.variable_scope('conv%d_%d' %
(num_filters, i + 1)):
conv = residual_block_deep_im(layers[-1],
num_filters, phase)
layers.append(conv)
assert conv.get_shape().as_list()[1:] ==
[56, 56, num_filters]
...](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-16-320.jpg)
![self.data_examples =
tf.placeholder(dtype=tf.float32,
shape=[batch_size,img_size,img_size,3])
self.data_labels =
tf.placeholder(dtype=tf.float32,
shape=[batch_size, num_labels])
self.queue =
tf.RandomShuffleQueue(shapes=[[img_size,
img_size, 3],[num_labels,]],
dtypes=[tf.float32, tf.float32],
capacity=capacity, min_after_dequeue=0)
self.enqueue_op =
self.queue.enqueue_many([self.data_examples,
self.data_labels)](https://image.slidesharecdn.com/random-170526125934/85/DataScience-Lab-2017_-TensorFlow_-18-320.jpg)
DataScience Lab 2017_Мониторинг модных трендов с помощью глубокого обучения и TensorFlow_Ольга Романюк
- 4. Kaiming He, Xiangyu Zhang, Shaoqing Ren, & Jian Sun. “Deep Residual Learning for Image Recognition”. arXiv 2015
- 5. Kaiming He, Xiangyu Zhang, Shaoqing Ren, & Jian Sun. “Deep Residual Learning for Image Recognition”. arXiv 2015
- 6. Kaiming He, Xiangyu Zhang, Shaoqing Ren, & Jian Sun. “Identity Mappings in Deep Residual Networks”.arXiv 2015
- 8. • • • • • •
- 9. graph = tf.Graph() with graph.as_default(): x = tf.placeholder(tf.float32,name='input') y = tf.placeholder(tf.float32,name='labels') x_image = tf.reshape(x, [-1, 28, 28, 1]) ...
- 10. with tf.variable_scope('conv1'): W_conv1 = weight_variable( [5, 5, 1, 32]) b_conv1 = bias_variable([32]) h_conv1 = tf.nn.relu( conv2d(x_image, W_conv1) + b_conv1) h_pool1 = max_pool_2x2(h_conv1) with tf.variable_scope('conv2'): ...
- 11. cross_entropy = tf.reduce_mean( tf.nn.softmax_cross_entropy_with_logits(labels=y,logits=y_conv)) train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy) correct_prediction = tf.equal(tf.argmax(y_conv, 1), tf.argmax(y, 1)) accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
- 12. with tf.Session(graph=graph) as sess: from tf.examples.tutorials.mnist import input_data mnist = input_data.read_data_sets('MNIST_data', one_hot=True) writer = tf.summary.FileWriter(logdir, sess.graph) sess.run(tf.global_variables_initializer()) for i in range(20000): batch = mnist.train.next_batch(50) _,train_accuracy = sess.run([train_step,accuracy], feed_dict={x: batch[0], y: batch[1]}) if i % 100 == 0: print("step %d, training accuracy %g" % (i, train_accuracy)) print("test accuracy %g" % sess.run(accuracy, feed_dict={x: mnist.test.images, y: mnist.test.labels}))
- 14. def conv_layer_resnet_im(inpt, filter_shape, stride, phase,name=''): filter_ = weight_variable(filter_shape, name=name + '_weights') normalized = batch_norm(inpt, phase,name=name) activated = tf.nn.relu(normalized) conv = tf.nn.conv2d(activated, filter=filter_, strides=[1, stride, stride, 1], padding="SAME") return conv
- 15. def residual_block_im(inpt, output_depth,phase, name=''): input_depth = inpt.get_shape().as_list()[3] conv1 = conv_layer_resnet_im(inpt, [3, 3,input_depth, output_depth], 1,phase, name=name+'_conv1') conv2 = conv_layer_resnet_im(conv1, [3, 3, output_depth, output_depth], 1,phase, name=name+'_conv2') if input_depth != output_depth: input_layer = tf.pad(inpt, [[0,0], [0,0], [0,0], [0, output_depth - input_depth]]) else: input_layer = inpt return conv2 + input_layer
- 16. num_blocks = 4 num_filters = 128 for i in range(num_blocks): with tf.variable_scope('conv%d_%d' % (num_filters, i + 1)): conv = residual_block_deep_im(layers[-1], num_filters, phase) layers.append(conv) assert conv.get_shape().as_list()[1:] == [56, 56, num_filters] ...
- 17. • • •
- 18. self.data_examples = tf.placeholder(dtype=tf.float32, shape=[batch_size,img_size,img_size,3]) self.data_labels = tf.placeholder(dtype=tf.float32, shape=[batch_size, num_labels]) self.queue = tf.RandomShuffleQueue(shapes=[[img_size, img_size, 3],[num_labels,]], dtypes=[tf.float32, tf.float32], capacity=capacity, min_after_dequeue=0) self.enqueue_op = self.queue.enqueue_many([self.data_examples, self.data_labels)
- 19. def put_inputs(self, sess): for data_examples, data_labels in self.iterator(...): sess.run(self.enqueue_op,feed_dict={ self.data_examples:data_examples, self.data_labels:data_labels}) def get_inputs(self): return self.queue.dequeue_many(self.batch_size)
- 21. for i in xrange(num_GPU): with tf.device('/gpu:%d' % i): with tf.name_scope('%s_%d' % ('GPU', i)) as name_scope: train_images_batch, train_labels_batch = train_runner.get_inputs() tower_loss = get_tower_loss(name_scope, train_images_batch, train_labels_batch,phase_train) tf.get_variable_scope().reuse_variables() loss_list.append(tower_loss) grads = optimizer.compute_gradients(tower_loss) tower_grads.append(grads)
- 22. grads = average_gradients(tower_grads) loss = tf.reduce_mean(loss_list) with tf.variable_scope(tf.get_variable_scope(), reuse=False): train_op = optimizer.apply_gradients(grads, global_step=global_step)
- 24. jeans