def generate_hashcode():
short_hash_list = []
train, train_label = get_data.get_zacao_input(train_dir, label_dir)
train_batch, train_label_batch = get_data.get_batch2(
train, train_label, 1, CAPACITY)
image_contents = tf.read_file(train_batch[0])
image_jpg = tf.image.decode_jpeg(image_contents, channels=3)
image_jpg = tf.image.resize_images(image_jpg, [IMG_W, IMG_H],
method=tf.image.ResizeMethod.AREA)
image_jpg = tf.reshape(tensor=image_jpg, shape=[1, IMG_W, IMG_H, 3])
logit, h_fc1_drop, hashcode = model.inference_without_dropout(
image_jpg, 1, N_CLASSES)
sess = tf.Session()
saver = tf.train.Saver()
print("Reading checkpoints...")
ckpt = tf.train.get_checkpoint_state(logs_train_dir)
if ckpt and ckpt.model_checkpoint_path:
global_step = ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
saver.restore(sess, ckpt.model_checkpoint_path)
print('Loading success, global_step is %s' % global_step)
else:
print('No checkpoint file found')
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
try:
step = 0
while step < len(train):
if coord.should_stop():
break
h_fc1_drop2, hashcode2, train_batch2, train_label_batch2 = sess.run(
[h_fc1_drop, hashcode, train_batch, train_label_batch])
print(h_fc1_drop2.shape)
print(h_fc1_drop2)
print(hashcode2.shape)
print(hashcode2)
print(train_label_batch2)
print(train_batch2)
img_name = train_batch2[0].decode()
short_hash_str = ','.join(str(i) for i in hashcode2[0])
long_hash_str = ','.join(
str(round(float(i), 8)) for i in h_fc1_drop2[0])
train_label_batch_str = str(train_label_batch2[0])
if short_hash_str not in short_hash_list:
short_hash_list.append(short_hash_str)
save_to_mysql2('suoyin', short_hash_str)
save_to_mysql('hash', long_hash_str, short_hash_str, img_name,
train_label_batch_str)
print('>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<')
step += 1
except tf.errors.OutOfRangeError:
print('Done training -- epoch limit reached')
finally:
coord.request_stop()
coord.join(threads)
sess.close()
def run_training():
# if tf.gfile.Exists(logs_train_dir):
# print(logs_train_dir)
# tf.gfile.DeleteRecursively(logs_train_dir)
# tf.gfile.MakeDirs(logs_train_dir)
train, train_label = get_data.get_zacao_input(train_dir, label_dir)
train_batch, train_label_batch = get_data.get_batch_jpg(train, train_label, IMG_W, IMG_H, BATCH_SIZE, CAPACITY)
train_logits = pre_model.inference(train_batch, BATCH_SIZE, N_CLASSES)
train_loss = pre_model.losses(train_logits, train_label_batch)
train_op = pre_model.trainning(train_loss, learning_rate)
train__acc = pre_model.evaluation(train_logits, train_label_batch)
summary_op = tf.summary.merge_all()
sess = tf.Session()
train_writer = tf.summary.FileWriter(logs_train_dir, sess.graph)
saver = tf.train.Saver()
sess.run(tf.global_variables_initializer())
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
step = 0
ckpt = tf.train.get_checkpoint_state(logs_train_dir)
if ckpt and ckpt.model_checkpoint_path:
step = int(ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1])
saver.restore(sess, ckpt.model_checkpoint_path)
print('Loading success, global_step is %d' % step)
step += 1
try:
while step < MAX_STEP:
if coord.should_stop():
break
_, tra_loss, tra_acc = sess.run([train_op, train_loss, train__acc])
if step % 50 == 0 or step + 1 == MAX_STEP:
print('Step %d, train loss = %.2f, train accuracy = %.2f%%' % (step, tra_loss, tra_acc * 100.0))
summary_str = sess.run(summary_op)
train_writer.add_summary(summary_str, step)
if step % 1000 == 0 or step + 1 == MAX_STEP:
checkpoint_path = os.path.join(logs_train_dir, 'model')
saver.save(sess, checkpoint_path, global_step=step)
step += 1
except tf.errors.OutOfRangeError:
print('Done training -- epoch limit reached')
finally:
coord.request_stop()
coord.join(threads)
sess.close()
def start_test():
with tf.Graph().as_default():
test, test_label = get_data.get_zacao_input(test_dir, test_label_dir)
test_batch, test_label_batch = get_data.get_batch_jpg(
test, test_label, IMG_W, IMG_H, BATCH_SIZE, CAPACITY)
test_logits, _, _ = model.inference_without_dropout(
test_batch, BATCH_SIZE, N_CLASSES)
test_acc = model.evaluation(test_logits, test_label_batch)
num_test = 1000
num_iter = int(math.ceil(num_test / BATCH_SIZE))
step = 0
saver = tf.train.Saver()
with tf.Session() as sess:
print("Reading checkpoints...")
ckpt = tf.train.get_checkpoint_state(logs_train_dir)
if ckpt and ckpt.model_checkpoint_path:
global_step = ckpt.model_checkpoint_path.split('/')[-1].split(
'-')[-1]
saver.restore(sess, ckpt.model_checkpoint_path)
print('Loading success, global_step is %s' % global_step)
else:
print('No checkpoint file found')
coord2 = tf.train.Coordinator()
threads2 = tf.train.start_queue_runners(sess=sess, coord=coord2)
try:
count_acc = 0
while step < num_iter:
if coord2.should_stop():
break
a, t = sess.run([test_batch, test_acc])
if step == 0:
print('Start test...')
count_acc += t
step += 1
print('Test acc = %.2f%%' % (count_acc / step * 100))
except tf.errors.OutOfRangeError:
print('Done testing -- epoch limit reached')
finally:
coord2.request_stop()
coord2.join(threads2)