官方Objection detection出问题

训练神经网络时，需要将标注好的数据转换成tensorflow特定数据，tfrecord。但是在转换时，发生以下错误：ValueError:Unicode strings with encoding declaration are not supported.Please use bytes or XML fragments without declaration.

 

  File "create_pascal_tf_record.py", line 185, in <module>
    tf.app.run()
  File "/home/chenyu/.local/lib/python3.5/site-packages/tensorflow/python/platform/app.py", line 125, in run
    _sys.exit(main(argv))
  File "create_pascal_tf_record.py", line 174, in main
    xml = etree.fromstring(xml_str)
  File "src/lxml/lxml.etree.pyx", line 3213, in lxml.etree.fromstring (src/lxml/lxml.etree.c:82934)
  File "src/lxml/parser.pxi", line 1814, in lxml.etree._parseMemoryDocument (src/lxml/lxml.etree.c:124471)
ValueError: Unicode strings with encoding declaration are not supported. Please use bytes input or XML fragments without declaration.
 

 

 

错误显示不支持的解码格式，以为时标注的xml文件出错了，就把相应的图片和标注文件删除了，发现还是出错。很感谢这篇博客的作者：https://blog.csdn.net/Fkk921912333/article/details/78537726 ，作者博客“解析 XML 字符串”部分，介绍了相关的xml文件解析方法，最主要的一句：print(etree.tostring(root, pretty_print=True).decode('utf-8'))。对比自己的creat_te_record文件，修改了读取文件时格式，即加入‘utf-8’，改变读取xml文件的编码方式，即可顺利转换数据。因为用的是tensorflow官方给的转换文件，具体语句为将xml = etree.fromstring(xml_str)改为xml = etree.fromstring(xml_str.encode('utf-8'))。可以据此更改自己的文件语句