Tensorflow 1.X basics

Here are the basics steps necessary to write TensorFlow code.

Note that this is for TensorFlow 1.X

Define constants

a = tf.constant(5.0)

Define variables

x = tf.Variable(init_value)

Define placeholders

p = tf.placeholder(tf.float32)

Define operations

c = x * x

Define way to measure loss

loss=tf.reduce_mean(tf.square(output - y))

Define optimizer

optimizer = tf.train.AdamOptimizer(learning_rate)

Apply optimizer to loss function

train = optimizer.minimize(loss)

init variables

init = tf.globalvariablesinitializer()

Run everything in a session

with tf.Session() as sess:
	sess.run(init)
	
	for epoch in range(epoch_count):
		sess.run(train, feed_dict={<feedDictHere})