Uncomment lines from this notebook to learn more.
In [1]:
# 1.1 Import tensorflow and other libraries.
import tensorflow as tf
import numpy as np
%matplotlib inline
import pylab
In [2]:
# 1.2 Create input data using NumPy. y = x * 0.1 + 0.3 + noise
x_train = np.random.rand(100).astype(np.float32)
noise = np.random.normal(scale=0.01, size=len(x_train))
y_train = x_train * 0.1 + 0.3 + noise
# Uncomment the following line to plot our input data.
pylab.plot(x_train, y_train, '.')
Out[2]:
In [3]:
# Create some fake evaluation data
x_eval = np.random.rand(len(x_train)).astype(np.float32)
noise = np.random.normal(scale=0.01, size=len(x_train))
y_eval = x_eval * 0.1 + 0.3 + noise
In [4]:
# 1.3 Buld inference graph.
# Create Variables W and b that compute y_data = W * x_data + b
W = tf.get_variable(shape=[], name='weights')
b = tf.get_variable(shape=[], name='bias')
# Uncomment the following lines to see W and b are.
# print(W)
# print(b)
# Create a placeholder we'll use later to feed x's into the graph for training and eval.
# shape=[None] means we can put in any number of examples.
# This is used for minibatch training, and to evaluate a lot of examples at once.
x = tf.placeholder(shape=[None], dtype=tf.float32, name='x')
# Uncomment this line to see what x is
# print(x)
# This is the same as tf.add(tf.mul(W, x), b), but looks nicer
y = W * x + b
In [7]:
# Write the graph so we can look at it in TensorBoard
# Now is a good time to try that
sw = tf.summary.FileWriter('/root/tensorboard/', graph=tf.get_default_graph())
In [14]:
# Create a placeholder we'll use later to feed the correct y value into the graph
y_label = tf.placeholder(shape=[None], dtype=tf.float32, name='y_label')
# print (y_label)
In [15]:
# 1.4 Build training graph.
loss = tf.reduce_mean(tf.square(y - y_label)) # Create an operation that calculates loss.
optimizer = tf.train.GradientDescentOptimizer(0.5) # Create an optimizer.
train = optimizer.minimize(loss) # Create an operation that minimizes loss.
# Uncomment the following 3 lines to see what 'loss', 'optimizer' and 'train' are.
# print("loss:", loss)
# print("optimizer:", optimizer)
# print("train:", train)
In [16]:
# Create an operation to initialize all the variables.
init = tf.global_variables_initializer()
# print(init)
In [17]:
# 1.6 Create a session and launch the graph.
sess = tf.Session()
sess.run(init)
# Uncomment the following line to see the initial W and b values.
# print(sess.run([W, b]))
In [18]:
# Uncomment these lines to test that we can compute a y from an x (without having trained anything).
# x must be a vector, hence [3] not just 3.
# x_in = [3]
# sess.run(y, feed_dict={x: x_in})
In [19]:
# Calculate accuracy on the evaluation data before training
def eval():
return sess.run(loss, feed_dict={x: x_eval, y_label: y_eval})
eval()
Out[19]:
In [35]:
# Add a summary so we can visualize the loss in TensorBoard
tf.summary.scalar('loss', loss)
summary_op = tf.summary.merge
In [36]:
# 1.7 Perform training.
for step in range(201):
# Run the training op; feed the training data into the graph
summary_str, _ = sess.run([summary_op, train], feed_dict={x: x_train, y_label: y_train})
sw.add_summary(summary_str, step)
# Uncomment the following two lines to watch training happen real time.
#if step % 20 == 0:
#print(step, sess.run([W, b]))
In [37]:
# 1.8 Uncomment the following lines to plot the predicted values
# pylab.plot(x_train, y_train, '.', label="target")
# pylab.plot(x_train, sess.run(y, feed_dict={x: x_train, y_label: y_train}), ".", label="predicted")
# pylab.legend()
# pylab.ylim(0, 1.0)
In [38]:
# Check accuracy on eval data after training
eval()
Out[38]:
Demonstrate saving and restoring a model
In [39]:
def predict(x_in): return sess.run(y, feed_dict={x: [x_in]})
In [45]:
# Save the model
saver = tf.train.Saver()
saver.save(sess, './my_checkpoint.ckpt')
Out[45]:
In [46]:
# Current prediction
predict(3)
Out[46]:
In [47]:
# Reset the model by running the init op again
sess.run(init)
In [48]:
# Prediction after variables reinitialized
predict(3)
Out[48]:
In [49]:
saver.restore(sess, './my_checkpoint.ckpt')
In [50]:
# Predictions after variables restored
predict(3)
Out[50]:
In [ ]: