def tower_loss(scope, images, labels):
"""Calculate the total loss on a single tower running the CIFAR model.
Args:
scope: unique prefix string identifying the CIFAR tower, e.g. 'tower_0'
images: Images. 4D tensor of shape [batch_size, height, width, 3].
labels: Labels. 1D tensor of shape [batch_size].
Returns:
Tensor of shape [] containing the total loss for a batch of data
"""
# Build inference Graph.
logits = inference.inference(images)
# Build the portion of the Graph calculating the losses. Note that we will
# assemble the total_loss using a custom function below.
_ = inference.loss(logits, labels)
# Assemble all of the losses for the current tower only.
losses = tf.get_collection('losses', scope)
# Calculate the total loss for the current tower.
total_loss = tf.add_n(losses, name='total_loss')
# Attach a scalar summary to all individual losses and the total loss; do the
# same for the averaged version of the losses.
for l in losses + [total_loss]:
# Remove 'tower_[0-9]/' from the name in case this is a multi-GPU training
# session. This helps the clarity of presentation on tensorboard.
loss_name = re.sub('%s_[0-9]*/' % inference.TOWER_NAME, '', l.op.name)
tf.summary.scalar(loss_name, l)
return total_loss
def train():
"""Train CIFAR-10 for a number of steps."""
with tf.Graph().as_default():
global_step = tf.contrib.framework.get_or_create_global_step()
# Get images and labels for CIFAR-10.
# Force input pipeline to CPU:0 to avoid operations sometimes ending up on
# GPU and resulting in a slow down.
with tf.device('/cpu:0'):
images, labels = inference.distorted_inputs()
# Build a Graph that computes the logits predictions from the
# inference model.
logits = inference.inference(images)
# Calculate loss.
loss = inference.loss(logits, labels)
# Build a Graph that trains the model with one batch of examples and
# updates the model parameters.
train_op = inference.train(loss, global_step)
class _LoggerHook(tf.train.SessionRunHook):
"""Logs loss and runtime."""
def begin(self):
self._step = -1
self._start_time = time.time()
def before_run(self, run_context):
self._step += 1
return tf.train.SessionRunArgs(loss) # Asks for loss value.
def after_run(self, run_context, run_values):
if self._step % FLAGS.log_frequency == 0:
current_time = time.time()
duration = current_time - self._start_time
self._start_time = current_time
loss_value = run_values.results
examples_per_sec = FLAGS.log_frequency * FLAGS.batch_size / duration
sec_per_batch = float(duration / FLAGS.log_frequency)
format_str = (
'%s: step %d, loss = %.2f (%.1f examples/sec; %.3f '
'sec/batch)')
print(format_str % (datetime.now(), self._step, loss_value,
examples_per_sec, sec_per_batch))
with tf.train.MonitoredTrainingSession(
checkpoint_dir=FLAGS.train_dir,
hooks=[
tf.train.StopAtStepHook(last_step=FLAGS.max_steps),
tf.train.NanTensorHook(loss),
_LoggerHook()
],
config=tf.ConfigProto(log_device_placement=FLAGS.
log_device_placement)) as mon_sess:
while not mon_sess.should_stop():
mon_sess.run(train_op)
def main(_):
Samples_placeholder = tf.placeholder(dtype=tf.float32,
shape=[None, 89],
name='X_input')
Labels_placeholder = tf.placeholder(dtype=tf.float32,
shape=[None, 5],
name='Y_input')
global_step = tf.Variable(0, trainable=False)
logits = inference.inference(Samples_placeholder)
loss = inference.loss(logits, Labels_placeholder)
train_op = inference.train(loss, global_step)
evaluation = inference.evaluation(logits, Labels_placeholder)
saver = tf.train.Saver()
with tf.Session() as sess:
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
model_dir = './model/train.ckpt-99999' #tf.train.latest_checkpoint('./logfile')
saver.restore(sess, model_dir)
test_data_file = os.path.join(FLAGS.data_dir, 'Data_o.mat')
if not tf.gfile.Exists(test_data_file):
raise ValueError('Failed to find file: ' + test_data_file)
data = scio.loadmat(test_data_file)
samples = np.array(data['Train_in'], dtype=np.float32)
labels = np.array(data['Train_out'], dtype=np.float32)
try:
eval_value, predict = sess.run([evaluation, logits],
feed_dict={
Samples_placeholder: samples,
Labels_placeholder: labels
})
print("Loss of testing NN")
print(eval_value)
test_file = os.path.join(FLAGS.data_dir, 'test_result.mat')
predict = predict.astype(np.int32)
scio.savemat(test_file, {'Predict': predict})
except tf.errors.OutOfRangeError:
print('Done testing --epoch limit reached')
finally:
coord.request_stop()
coord.join(threads)
def train():
"""
:return:
"""
# 读取数据集
filenames = os.listdir(datasets_dir)
# 过滤不合格数据集
for filename in filenames:
if not os.path.splitext(filename)[1] == '.pickle':
filenames.remove(filename)
logits = inference.inference(image_holder, reuse=False)
global_step = tf.Variable(0, trainable=False)
# 定义滑动平滑平均值
variable_averages = tf.train.ExponentialMovingAverage(
MOVING_AVERAGE, global_step)
variable_averages_op = variable_averages.apply(tf.trainable_variables())
# 损失函数值
loss = inference.loss(logits, label_holder)
# 使用反向传播函数之前优化学习率
learning_rate = tf.train.exponential_decay(LEARNING_RATE_BASE,
global_step,
MAX_STEPS,
decay_rate=LEARNING_RATE_DECAY)
# 定义反向传播函数
train_step = tf.train.GradientDescentOptimizer(learning_rate).minimize(
loss, global_step=global_step)
# 使用反向函数和滑动平滑值更新参数
train_op = tf.group(train_step, variable_averages_op)
saver = tf.train.Saver()
with tf.Session() as sess:
tf.global_variables_initializer().run()
tf.train.start_queue_runners()
if not os.path.exists(models_dir):
os.makedirs(models_dir)
for step in range(MAX_STEPS):
for filename in filenames:
train_image, train_label = dataset.read(filename)
assert isinstance(train_image, list)
assert isinstance(train_label, list)
_, loss_value = sess.run([train_op, loss],
feed_dict={
image_holder: train_image,
label_holder: train_label
})
if step % 2 == 0:
print("after %d steps, the loss value is %g" %
(step, loss_value))
saver.save(sess, models_file, global_step=step)
def main(_):
Samples_placeholder = tf.placeholder(dtype=tf.float32,shape=[None,89],name='X_input')
Labels_placeholder = tf.placeholder(dtype=tf.float32,shape=[None,5],name='Y_input')
global_step = tf.Variable(0,trainable=False)
logits = inference.inference(Samples_placeholder)
loss = inference.loss(logits,Labels_placeholder)
train_op = inference.train(loss,global_step)
evaluation = inference.evaluation(logits, Labels_placeholder)
saver = tf.train.Saver()
summary = tf.summary.merge_all()
init = tf.global_variables_initializer()
with tf.Session() as sess:
summary_writer = tf.summary.FileWriter(FLAGS.train_dir, sess.graph)
sess.run(init)
# load data
train_file = os.path.join(FLAGS.data_dir,'Data_o.mat' )
if not tf.gfile.Exists(train_file):
raise ValueError('Failed to find file: ' + train_file)
data = scio.loadmat(train_file)
samples = np.array(data['Train_data'],dtype=np.float32)
labels = np.array(data['Train_label'],dtype=np.float32)
train_samples = samples
train_labels = labels
test_samples = samples
test_labels = labels
#Para = data['Data_PS']
for i in xrange(FLAGS.max_steps):
start_time = time.time()
loss_value,_=sess.run([loss,train_op],feed_dict={Samples_placeholder:samples,Labels_placeholder:labels})
duration=time.time()-start_time
if (i+1) % 10 == 0:
print("Loss of training NN, step: %d, loss: %f, and time:%f" % (i+1,loss_value,duration))
#if (i+1) % 100 == 0 :
eval_value = sess.run(evaluation, feed_dict={Samples_placeholder:train_samples,Labels_placeholder:train_labels})
#print("Evaluation of test samples, step: %d, loss: %f." % (i+1,eval_value))
#summary_writer.add_summary(summary_str, i)
if i%1000 == 0 or (i+1) == FLAGS.max_steps:
eval_value = sess.run(evaluation, feed_dict={Samples_placeholder:samples,Labels_placeholder:labels})
print(eval_value)
checkpoint_path = os.path.join(FLAGS.train_dir, 'train.ckpt')
saver.save(sess, checkpoint_path, global_step=i)
predict = sess.run(logits, feed_dict={Samples_placeholder:test_samples,Labels_placeholder:test_labels})
test_file = os.path.join(FLAGS.data_dir,'predict.mat' )
scio.savemat(test_file,{'Predict':predict})
def main(_):
Samples_placeholder = tf.placeholder(dtype=tf.float32,
shape=[None, 89],
name='X_input')
Labels_placeholder = tf.placeholder(dtype=tf.float32,
shape=[None, 5],
name='Y_input')
global_step = tf.Variable(0, trainable=False)
logits = inference.inference(Samples_placeholder) #网络结构
loss = inference.loss(logits, Labels_placeholder) #误差函数
train_op = inference.train(loss, global_step) #训练方式
evaluation = inference.evaluation(logits,
Labels_placeholder) #预测结果和实际结果评估函数
saver = tf.train.Saver()
with tf.Session() as sess:
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
model_dir = './model/train.ckpt-99999' #tf.train.latest_checkpoint('./logfile')
saver.restore(sess, model_dir)
test_data_file = os.path.join(FLAGS.data_dir, 'Data_o.mat')
if not tf.gfile.Exists(test_data_file):
raise ValueError('Failed to find file: ' + test_data_file)
data = scio.loadmat(test_data_file)
#samples = np.array(data['Train_in'],dtype=np.float32)
labels = np.array(data['Train_out'], dtype=np.float32)
#新增
# Connect to the database
connection = pymysql.connect(
host='localhost',
user='******',
password='******',
db='weibodata',
charset='utf8mb4',
)
try:
# with connection.cursor() as cursor:
# Create a new record
# sql = "INSERT INTO `users` (`email`, `password`) VALUES (%s, %s)"
# cursor.execute(sql, ('*****@*****.**', 'very-secret'))
# connection is not autocommit by default. So you must commit to save
# your changes.
# connection.commit()
with connection.cursor() as cursor:
# Read a single record
sql = "SELECT * FROM `向量结果` WHERE `微博昵称`=%s"
cursor.execute(sql, (sys.argv[1], ))
#sql = "SELECT `微博昵称`, `严谨性` FROM `sheet1`ORDER BY `微博昵称` "
#cursor.execute(sql)
result = cursor.fetchone()
arr_ys = list(result)
index = 0
arr = []
for arr_y in arr_ys:
if (index not in [
102, 101, 100, 99, 93, 92, 89, 80, 77, 15, 14, 13,
10, 0
]):
arr.append(arr_y)
index += 1
#print(arr)
finally:
connection.close()
samples = np.array(arr, dtype=np.float32)
samples = samples.reshape((1, 89))
#新增结束
try:
eval_value, predict = sess.run([evaluation, logits],
feed_dict={
Samples_placeholder: samples,
Labels_placeholder: labels
}) #打印loss
print("Loss of testing NN")
print(eval_value)
test_file = os.path.join(FLAGS.data_dir, 'test_result.mat')
predict = predict.astype(np.int32)
scio.savemat(test_file, {'Predict': predict})
except tf.errors.OutOfRangeError:
print('Done testing --epoch limit reached')
finally:
coord.request_stop()
coord.join(threads)
