In [1]:

import numpy as np
import numpy.random as rng
import tensorflow as tf

In [2]:

n = 1000
din = np.random.randint(0,2, (n, 2)).astype(bool)
out = np.logical_xor(din[:,0], din[:,1])

out = np.zeros((n, 2), dtype="bool")

out[:,1] = np.logical_xor(din[:,0], din[:,1])
out[:,0] = True ^ out[:,1]

In [3]:

x = tf.placeholder("bool", shape=[None,2])
x1 = tf.cast(x, "float")

with tf.name_scope('Layer1'):
    W1 = tf.Variable(tf.random_normal([2,10]))
    b1 = tf.Variable(tf.random_normal([10]))
    hidden = tf.nn.relu(tf.matmul(x1, W1) + b1)
    
    
with tf.name_scope('Layer2'):
    W2 = tf.Variable(tf.random_normal([10,2]))
    b2 = tf.Variable(tf.random_normal([2]))
    hidden1 =  tf.matmul(hidden, W2) + b2

with tf.name_scope('output'):
    y = tf.nn.softmax(hidden1)
    y_val = tf.arg_max(y,1)

In [4]:

y_ = tf.placeholder("float",[None,2])

loss = tf.reduce_sum(-y_*tf.log(y))
init = tf.initialize_all_variables()
optimizer = tf.train.GradientDescentOptimizer(.001).minimize(loss)


tf.scalar_summary("cost", loss)

merged = tf.merge_all_summaries()

In [5]:

fd = {x: din, y_:out}
with tf.Session() as sess:
    sess.run(init)
    summary_writer = tf.train.SummaryWriter("./logs/run1")
    summary_writer.add_graph(y.graph)
    
    print(sess.run(loss, feed_dict=fd))
    
    
    
    for step in range(40):
        summ, cost_curr, _ = sess.run([merged, loss, optimizer], feed_dict={x: din, y_: out})
        if step % 10 == 0 : print("Iteration", step, "Loss", cost_curr)
            
        summary_writer.add_summary(summ, step)
            
            
#     print(sess.run(y_val, feed_dict={x:[[0,0], [1,1], [1,0], [0,1]]}))

3329.73
Iteration 0 Loss 3329.73
Iteration 10 Loss 277.95
Iteration 20 Loss 126.06
Iteration 30 Loss 72.0056

In [ ]: