목록Cloud Computing/Tensorflow (3)
MJay
conv1 만 프린트 하는 command print (tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope='conv1')) variable 읽는 코드 (sort팅) reader = tf.train.load_checkpoint("/tmp/mj/cifar10_train") variable_map = reader.get_variable_to_shape_map() names = (variable_map.keys()) result = [] for name in names: result.append((name, variable_map[name])) pprint.pprint(result) reader = tf.train.NewCheckpointReader(..
Graphs are sets of connected nodes - verticesThe connections are referred to as edgeseach node is an operation with possible inputs that can supply some outputIn [1]:import tensorflow as tf In [2]:n1 = tf.constant(1) In [3]:n2 = tf.constant(2) In [4]:n3 = n1 + n2 In [5]:with tf.Session() as sess: result = sess.run(n3) In [6]:print (result) 3 n3 은 여기서 뭘까 -> Tensor 이다In [7]:print (tf.get_default_g..