版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/wang15061955806/article/details/82114235
1、slim模块作为一种轻量级的tensorflow库,使得模型的构建,训练,测试都变得更加简单
使用时需要import tf.contrib.slim as slim
参考：https://www.2cto.com/kf/201706/649266.html
2、tf.global_variables_initializer 问题
python fcn_32s_test_pascal.py时出现 _, serialized_example = reader.read(tfrecord_filenames_queue)
OutOfRangeError (see above for traceback): FIFOQueue '_0_input_producer' is closed and has insufficient elements (requested 1, current size 0)
[[Node: ReaderReadV2 = ReaderReadV2[_device="/job:localhost/replica:0/task:0/device:CPU:0"](TFRecordReaderV2, input_producer)]]
网上说是需要将tf.global_variables_initializer转化为tf.local_variables_initializer，我尝试了下，就没有出现这个错误了。
3、Numpy操作函数
print((x + y)[..., newaxis]) #提高维度
imsize = 3
x, y = ogrid[:imsize, :imsize] 获取2个数组（3,1）（1,3）