def predict(image_path):
size = FLAGS.image_size, FLAGS.image_size
img = get_X(image_path, size)
ret = model.predict(np.array([img]))
ret = np.squeeze(ret)
index = np.argmax(ret, axis=0)
text = int2text(index, FLAGS.wordset)
return text
def predict(image_path):
size = FLAGS.image_width, FLAGS.image_height
img = get_X(image_path, size)
ret = model.predict(np.array([img]))
ret = ret.reshape(FLAGS.captcha_size, FLAGS.charset_size)
index = np.argmax(ret, axis=1)
text = index2text(index, FLAGS.charset)
return text
def predict_top_5(image_path):
size = FLAGS.image_size, FLAGS.image_size
img = get_X(image_path, size)
ret = model.predict(np.array([img]))
ret = np.squeeze(ret)
ret = np.argsort(ret)
ret = ret[-5:][::-1]
text = [int2text(index, FLAGS.wordset) for index in ret]
return text
def predict(img):
size = (FLAGS.image_size, FLAGS.image_size)
image = get_X(img, size)
predict_opts = [graph['predicted_val_top_k'], graph['predicted_index_top_k']]
feed_dict = {graph['images']: [image],
graph['keep_prob']: 1.0,
graph['is_training']: False}
predict_val, predict_index = sess.run(predict_opts, feed_dict=feed_dict)
predict_index = predict_index.flatten()
text = int2text(predict_index[0], FLAGS.wordset)
return text
def inference(image):
print('inference')
from preprocess import get_X
temp_image = get_X(image)
with tf.Session() as sess:
logger.info('========start inference============')
# images = tf.placeholder(dtype=tf.float32, shape=[None, 64, 64, 1])
# Pass a shadow label 0. This label will not affect the computation graph.
graph = build_graph(top_k=3)
saver = tf.train.Saver()
ckpt = tf.train.latest_checkpoint(FLAGS.checkpoint_dir)
if ckpt:
saver.restore(sess, ckpt)
predict_val, predict_index = sess.run(
[graph['predicted_val_top_k'], graph['predicted_index_top_k']],
feed_dict={
graph['images']: temp_image,
graph['keep_prob']: 1.0,
graph['is_training']: False
})
return predict_val, predict_index
import pandas as pd
import preprocess
import tensorflow as tf
from sklearn.preprocessing import MultiLabelBinarizer
from sklearn.utils import shuffle
data = pd.read_csv("myFP_217_D2.csv", header=None)
D2 = preprocess.get_data(data)
X = preprocess.get_X(D2)
X = preprocess.comb(X)
# y = pd.DataFrame(preprocess.get_target(D2))
value = preprocess.get_target(D2)
value = MultiLabelBinarizer().fit_transform(value)
y = pd.DataFrame(value)
X, y = shuffle(X, y, random_state=0)
X_train, X_test = X[:int((0.8 * len(X)))], X[int((0.8 * len(X))):]
y_train, y_test = y[:int((0.8 * len(X)))], y[int((0.8 * len(X))):]
def rnn1():
n_nodes_hl1 = 1024
n_nodes_hl2 = 512
n_nodes_hl3 = 256
n_classes = 7
batch_size = 100
x = tf.placeholder(tf.float32, [None, 2048], name="features")