标签：img 300 用法 a1 维度 split tf
from PIL import Image
import numpy as np
import tensorflow as tf
'''
split 对维度进行分割
tf.split(
data, 数据图片 ( 300*600*3)
num_or_size_splits , 分割的数组 传个数
axis， 代表维度，当前的维度为 0 1 2
)
tf.split(data,3,2) 得到数据的维度为 [(300,600,1),(300,600,1),(300,600,1)]
tf.split(data,[100.200,300],1) 得到数据的维度为 [(300,100,3),(300,100,3),(300,200,3)]
'''
img = Image.open('./test_data/tabby_cat.png')
img = np.array(img)
a = tf.split(img,3,2)
with tf.Session() as sess:
sess.run(tf.global_variables_initializer())
a1 = sess.run([a ])
# for i in range(len(a1[0])):
# print(a1[i].shape )
for i in range(len(a1[0])):
# print(type(a1[0][i]))
print(a1[0][i].shape)
# print(img.shape)
标签：img,
300,
用法,
a1,
维度,
split,
tf
From： https://blog.51cto.com/u_15955675/6039837