def test(mnist):
with tf.Graph().as_default() as graph:
x = tf.placeholder(tf.float32, shape = [None, rnn_forward.TIME_STEPS, rnn_forward.INPUT_ROWS])
y_ = tf.placeholder(tf.float32, shape = [None, rnn_forward.OUTPUT_NODE])
y = rnn_forward.forward(x,rnn_backward.REGULARIZER)
ema = tf.train.ExponentialMovingAverage(rnn_backward.MOVING_AVERAGE_DECAY)
ema_restore = ema.variables_to_restore()
saver = tf.train.Saver(ema_restore)
correct = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))
accuracy = tf.reduce_mean(tf.cast(correct, tf.float32))
while True:
with tf.Session() as sess:
ckpt = tf.train.get_checkpoint_state(rnn_backward.MODEL_SAVE_PATH)
if ckpt and ckpt.model_checkpoint_path:
saver.restore(sess, ckpt.model_checkpoint_path)
global_step = ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
xs = mnist.test.images.reshape((-1, rnn_forward.TIME_STEPS, rnn_forward.INPUT_ROWS))
accuracy_score = sess.run(accuracy, feed_dict = {x:xs, y_:mnist.test.labels})
print("after %s setps, test accuracy is %g"%(global_step, accuracy_score))
else:
print("no checkpoint")
return
time.sleep(TEST_INTERVAL)
def backward(mnist):
x = tf.placeholder(tf.float32, shape = [None, rnn_forward.TIME_STEPS, rnn_forward.INPUT_ROWS])
y_ = tf.placeholder(tf.float32, shape = [None, rnn_forward.OUTPUT_NODE])
y = rnn_forward.forward(x,REGULARIZER)
global_step = tf.Variable(0, trainable = False)
learning_rate = tf.train.exponential_decay(
LEARNING_RATE_BASE,
global_step,
mnist.train.num_examples/BATCH_SIZE,
LEARNING_RATE_DECAY,
staircase = True)
ce = tf.nn.sparse_softmax_cross_entropy_with_logits(logits = y, labels = tf.argmax(y_, 1))
loss_ce = tf.reduce_mean(ce)
loss_total = loss_ce + tf.add_n(tf.get_collection('losses'))
train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(loss_total, global_step = global_step)
ema = tf.train.ExponentialMovingAverage(MOVING_AVERAGE_DECAY, global_step)
ema_op = ema.apply(tf.trainable_variables())
with tf.control_dependencies([train_step, ema_op]) :
train_op = tf.no_op(name = 'train')
saver = tf.train.Saver()
writer = tf.summary.FileWriter('W:\\python\\tensorflow\\file', tf.get_default_graph())
writer.close()
with tf.Session() as sess:
init_op = tf.global_variables_initializer()
sess.run(init_op)
for i in range(STEPS):
xs, ys = mnist.train.next_batch(BATCH_SIZE)
xs = xs.reshape((BATCH_SIZE, rnn_forward.TIME_STEPS, rnn_forward.INPUT_ROWS))
_, loss_value, step = sess.run([train_op, loss_total, global_step], feed_dict = {x: xs, y_: ys})
if i % 500 == 0:
print("After %d steps, loss is: %f" %(step, loss_value))
saver.save(sess, os.path.join(MODEL_SAVE_PATH, MODEL_NAME), global_step = global_step)
def test(mnist):
with tf.Graph().as_default() as graph:
x = tf.placeholder(
tf.float32,
shape=[None, rnn_forward.TIME_STEPS, rnn_forward.INPUT_ROWS])
y_ = tf.placeholder(tf.float32, shape=[None, rnn_forward.OUTPUT_NODE])
y = rnn_forward.forward(x, rnn_backward.REGULARIZER)
ema = tf.train.ExponentialMovingAverage(
rnn_backward.MOVING_AVERAGE_DECAY)
ema_restore = ema.variables_to_restore()
saver = tf.train.Saver(ema_restore)
correct = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))
accuracy = tf.reduce_mean(tf.cast(correct, tf.float32))
while True:
with tf.Session() as sess:
ckpt = tf.train.get_checkpoint_state(
rnn_backward.MODEL_SAVE_PATH)
if ckpt and ckpt.model_checkpoint_path:
saver.restore(sess, ckpt.model_checkpoint_path)
global_step = ckpt.model_checkpoint_path.split(
'/')[-1].split('-')[-1]
xs = mnist.test.images.reshape(
(-1, rnn_forward.TIME_STEPS, rnn_forward.INPUT_ROWS))
accuracy_score = sess.run(accuracy,
feed_dict={
x: xs,
y_: mnist.test.labels
})
print("after %s setps, test accuracy is %g" %
(global_step, accuracy_score))
else:
print("no checkpoint")
return
time.sleep(TEST_INTERVAL)
from rnn_forward import forward
forward('82', 2)
forward('90', 4)
forward('98', 3)
forward('11', 3)
from rnn_forward import forward
forward('51', 2)
forward('59', 5)
forward('67', 1)
forward('75', 0)
from rnn_forward import forward
forward('0',4)
forward('27',0)
forward('35',1)
forward('43',5)