tensorflow1.6 繼續訓練操作

sess=tf.Session()

sess.run(tf.global_variables_initializer)

ckpt= tf.train.get_checkpoint_state(check_point_path)#這裏的check_point_path是checkpoint所在的文件夾路徑

saver=tf.train.Saver()

if ckpt and ckpt.model_checkpoint_path:

    saver.restore(sess,ckpt.model_checkpoint_path)

train...