correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1))162-2correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(y_,1))
