NiftyNet net_segment没有属性“_logger”

Question

当我尝试对从PyPI安装的NiftyNet执行快速入门说明时，在运行net_segment时得到以下错误：

net_segment inference -c ~/niftynet/extensions/dense_vnet_abdominal_ct/config.ini 
WARNING:tensorflow:SimpleITK adapter failed to load, reducing the supported file formats.
Traceback (most recent call last):
  File "/usr/local/bin/net_segment", line 11, in <module>
    sys.exit(main())
  File "/usr/local/lib/python2.7/dist-packages/niftynet/__init__.py", line 51, in main
    set_logger()
  File "/usr/local/lib/python2.7/dist-packages/niftynet/io/misc_io.py", line 585, in set_logger
    tf.logging._logger.handlers = []
AttributeError: 'module' object has no attribute '_logger'

我已经在Linux Mint18.3安装(这个)以及带有SimpleITK的Arch Linux安装上尝试过了。

在安装了SimpleITK的AUR上安装了NiftyNet的Arch Linux上也出现了同样的错误。

net_segment inference -c ~/niftynet/extensions/dense_vnet_abdominal_ct/config.ini
Traceback (most recent call last):
  File "/usr/bin/net_segment", line 11, in <module>
    load_entry_point('NiftyNet==0.2.2', 'console_scripts', 'net_segment')()
  File "/usr/lib/python3.6/site-packages/niftynet/__init__.py", line 51, in main
    set_logger()
  File "/usr/lib/python3.6/site-packages/niftynet/io/misc_io.py", line 585, in set_logger
    tf.logging._logger.handlers = []
AttributeError: module 'tensorflow.tools.api.generator.api.logging' has no attribute '_logger'

通过将TensorFlow降级到版本1.7解决了此问题。

Answer 1

正如您已经发现的，NiftyNet还不支持TF 1.8。