Downloaded 36 times
![Perceptron classifier for
MNIST 2/2
import tensorflow as tf
x = tf.placeholder(tf.float32, [None, 784])
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))
y = tf.nn.softmax(tf.matmul(x, W) + b)
y_ = tf.placeholder(tf.float32, [None, 10])
cross_entropy = -tf.reduce_sum(y_*tf.log(y))
train_step =
tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)
for i in range(1000):
batch_xs, batch_ys = mnist.train.next_batch(100)
sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})](https://image.slidesharecdn.com/tensorflowppt-160408142819/75/TensorFlow-6-2048.jpg)
![Convolution for MNIST
#Convolutional Layer
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)
#Densely Connected Layer
W_fc1 = weight_variable([12 * 12 * 64, 1024])
b_fc1 = bias_variable([1024])
h_pool1_flat = tf.reshape(h_pool1, [-1, 12*12*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool1_flat, W_fc1) + b_fc1)
#Dropout
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)](https://image.slidesharecdn.com/tensorflowppt-160408142819/75/TensorFlow-9-2048.jpg)
![Language Modeling with
LSTM
lstm = rnn_cell.BasicLSTMCell(lstm_size)
stacked_lstm = rnn_cell.MultiRNNCell([lstm] * number_of_layers)
# Initial state of the LSTM memory.
state = tf.zeros([batch_size, lstm_size])
loss = 0.0
for current_batch_of_chars in dataset:
# State is updated after processing each batch of chars.
output, state = stacked_lstm(current_batch_of_chars, state)
# The LSTM output can be used to make next word predictions
logits = tf.matmul(output, softmax_w) + softmax_b
probabilities = tf.nn.softmax(logits)
loss += loss_function(probabilities, target_words)](https://image.slidesharecdn.com/tensorflowppt-160408142819/75/TensorFlow-12-2048.jpg)
![Perceptron classifier for
MNIST 2/2
import tensorflow as tf
x = tf.placeholder(tf.float32, [None, 784])
W = tf.Variable(tf.zeros([784, 10]))
b = tf.Variable(tf.zeros([10]))
y = tf.nn.softmax(tf.matmul(x, W) + b)
y_ = tf.placeholder(tf.float32, [None, 10])
cross_entropy = -tf.reduce_sum(y_*tf.log(y))
train_step =
tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy)
for i in range(1000):
batch_xs, batch_ys = mnist.train.next_batch(100)
sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})](https://crownmelresort.com/image.slidesharecdn.com/tensorflowppt-160408142819/75/TensorFlow-6-2048.jpg)
![Convolution for MNIST
#Convolutional Layer
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)
#Densely Connected Layer
W_fc1 = weight_variable([12 * 12 * 64, 1024])
b_fc1 = bias_variable([1024])
h_pool1_flat = tf.reshape(h_pool1, [-1, 12*12*64])
h_fc1 = tf.nn.relu(tf.matmul(h_pool1_flat, W_fc1) + b_fc1)
#Dropout
keep_prob = tf.placeholder("float")
h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)](https://crownmelresort.com/image.slidesharecdn.com/tensorflowppt-160408142819/75/TensorFlow-9-2048.jpg)
![Language Modeling with
LSTM
lstm = rnn_cell.BasicLSTMCell(lstm_size)
stacked_lstm = rnn_cell.MultiRNNCell([lstm] * number_of_layers)
# Initial state of the LSTM memory.
state = tf.zeros([batch_size, lstm_size])
loss = 0.0
for current_batch_of_chars in dataset:
# State is updated after processing each batch of chars.
output, state = stacked_lstm(current_batch_of_chars, state)
# The LSTM output can be used to make next word predictions
logits = tf.matmul(output, softmax_w) + softmax_b
probabilities = tf.nn.softmax(logits)
loss += loss_function(probabilities, target_words)](https://crownmelresort.com/image.slidesharecdn.com/tensorflowppt-160408142819/75/TensorFlow-12-2048.jpg)
More Related Content
![[신경망기초] 합성곱신경망](https://cdn.slidesharecdn.com/ss_thumbnails/2-180604135842-thumbnail.jpg?width=640&height=640&fit=bounds)
![[신경망기초]오류역전파알고리즘구현](https://cdn.slidesharecdn.com/ss_thumbnails/nn-190319093725-thumbnail.jpg?width=640&height=640&fit=bounds)
TensorFlow
- 1. Jiří Materna Head ofResearch, Seznam.cz TensorFlow
- 2. What is notTensorFlow
- 3.
- 4.
- 5.
- 6. Perceptron classifier for MNIST2/2 import tensorflow as tf x = tf.placeholder(tf.float32, [None, 784]) W = tf.Variable(tf.zeros([784, 10])) b = tf.Variable(tf.zeros([10])) y = tf.nn.softmax(tf.matmul(x, W) + b) y_ = tf.placeholder(tf.float32, [None, 10]) cross_entropy = -tf.reduce_sum(y_*tf.log(y)) train_step = tf.train.GradientDescentOptimizer(0.01).minimize(cross_entropy) for i in range(1000): batch_xs, batch_ys = mnist.train.next_batch(100) sess.run(train_step, feed_dict={x: batch_xs, y_: batch_ys})
- 7.
- 8.
- 9. Convolution for MNIST #ConvolutionalLayer 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) #Densely Connected Layer W_fc1 = weight_variable([12 * 12 * 64, 1024]) b_fc1 = bias_variable([1024]) h_pool1_flat = tf.reshape(h_pool1, [-1, 12*12*64]) h_fc1 = tf.nn.relu(tf.matmul(h_pool1_flat, W_fc1) + b_fc1) #Dropout keep_prob = tf.placeholder("float") h_fc1_drop = tf.nn.dropout(h_fc1, keep_prob)
- 10.
- 11.
- 12. Language Modeling with LSTM lstm= rnn_cell.BasicLSTMCell(lstm_size) stacked_lstm = rnn_cell.MultiRNNCell([lstm] * number_of_layers) # Initial state of the LSTM memory. state = tf.zeros([batch_size, lstm_size]) loss = 0.0 for current_batch_of_chars in dataset: # State is updated after processing each batch of chars. output, state = stacked_lstm(current_batch_of_chars, state) # The LSTM output can be used to make next word predictions logits = tf.matmul(output, softmax_w) + softmax_b probabilities = tf.nn.softmax(logits) loss += loss_function(probabilities, target_words)
- 13.
- 14.
- 15. Try Keras from keras.modelsimport Sequential from keras.layers.core import Dense, Dropout, Activation, Flatten from keras.layers.convolutional import Convolution2D, MaxPooling2D from keras.optimizers import SGD model = Sequential() model.add(Convolution2D(32, 5, 5, border_mode=‘same', input_shape=(1, 28, 28))) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Dense(1024)) model.add(Flatten()) model.add(Activation('relu')) model.add(Dropout(0.5)) model.add(Activation('softmax')) sgd = SGD(lr=0.1) model.compile(loss='categorical_crossentropy', optimizer=sgd)
- 16.
- 17. Time to changeyour job
- 18.
- 19.