tensorflow 代码实战报错集合整理

1、在xxx.sh脚本里使用 python -O(大写英文O)  关闭 assert 断言 AssertError 报错

python -O 

到百度首页 百度首页登录 Python中使用assert语句需要注意什么

python 错误处理 assert

2、Debug 报下面错误：

Traceback (most recent call last):
  File "D:\Program Files\JetBrains\PyCharm\helpers\pydev\_pydevd_bundle\pydevd_comm.py", line 393, in _on_run
    r = r.decode('utf-8')
UnicodeDecodeError: 'utf-8' codec can't decode bytes in position 1022-1023: unexpected end of data

解决方案：

在pycharm中两个地方设置成utf-8，页面右下角和File>setting中编码设置成UTF-8

参考：pycharm中可以运行脚本但是不能debug脚本

pycharm debug出现UnicodeDecodeError: 'utf-8' codec can't decode 解决办法

第一处：

上图右下角被水印覆盖掉，去除的方式：

自定义CSDN博客中图片上的水印内容

下面是插入图片的时候csdn-markdown编辑器中生成的代码：

https://img-blog.csdnimg.cn/20190315101340322.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3hyaW5vc3ZpcA==,size_16,color_FFFFFF,t_70

注意：这里的参数用Base64加密算法

对上面的链接作仔细观察，发现有“？x-oss-process=image”字样。接下来尝试把代码中“？x-oss-process=image”后面的部分去掉， 即：

https://img-blog.csdnimg.cn/20190315101340322.png?

 如下图，便去掉了水印

第二处：

若上述方法还不行，试试下面的方法： 

3、训练 rsgan 报如下错误：

Traceback (most recent call last):
  File "E:/python/Tensorflow/随书资源/代码/12-9  rsgan.py", line 288, in <module>
    showresult(163,"near",imagesv,x_nearestv)
  File "E:/python/Tensorflow/随书资源/代码/12-9  rsgan.py", line 61, in showresult
    ypsnr=batch_y_psnr(conimg)
  File "E:/python/Tensorflow/随书资源/代码/12-9  rsgan.py", line 34, in batch_y_psnr
    psnr=np.mean(20*np.log10(255.0/np.sqrt(mse)))
UnboundLocalError: local variable 'mse' referenced before assignment

原因：代码第34行  局部变量 mse 先调用    后定义 导致报错

解决办法： 对调上图的两行代码

4、run out of memory 

Tensor Flow: Ran out of memory trying to allocate

Allocator (GPU_0_bfc) ran out of memory trying to allocate 2.14GiB. The caller indicates that this is not a failure, but may mean that there could be performance gains if more memory were available.

解决办法： 减小 batch_size

5、在tensorflow训练model时，添加 vgg_19 网络，报下面错误：

ValueError: Negative dimension size caused by subtracting 7 from 4 for 'vgg_19/fc6/Conv2D' (op: 'Conv2D') with input shapes: [48,4,2,512], [7,7,512,4096].

原因： vgg_19 的输入图片的大小应为  256x256x3 的图片，若出入的不是这个 shape 就会报这个错误 

解决办法： 通过 padding 或者 clip 将输入图片修改为 256x256x3 的图片

6  、AttributeError: 'dict' object has no attribute 'SerializeToString'

 参考： TensorFlow报错：'dict' object has no attribute 'SerializeToString'

Traceback (most recent call last):
  File "main.py", line 37, in <module>
    main(config)
  File "main.py", line 29, in main
    trainer.train()
  File "/home/hyw/zp/tf/Pose-Guided-Person-Image-Generation-master/trainer.py", line 397, in train
    self.generate(x, x_target, pose_target_fixed, self.model_dir, idx=step)
  File "/home/hyw/zp/tf/Pose-Guided-Person-Image-Generation-master/trainer.py", line 476, in generate
    G, G_1 = self.sess.run(self.G, self.G1, {self.x: x_fixed, self.pose_target: pose_target_fixed})
  File "/home/hyw/zp/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 884, in run
    compat.as_bytes(options.SerializeToString())) if options else None
AttributeError: 'dict' object has no attribute 'SerializeToString'

原因： 这个报错是因为sess.run（a，b，c）里要运行的多个变量没有用中括号括起来

解决办法： 把run() 要运行的多个变量用中括号括起来，即 sess.run([a,b,c], {x:x,y:y}

7、IOError: CRC check failed 0x49a91344 != 0x6adf6a7L

Traceback (most recent call last):
  File "score.py", line 13, in <module>
    import tflib.inception_score
  File "/home/hyw/zp/tf/Pose-Guided-Person-Image-Generation-master/tflib/inception_score.py", line 101, in <module>
    _init_inception()
  File "/home/hyw/zp/tf/Pose-Guided-Person-Image-Generation-master/tflib/inception_score.py", line 73, in _init_inception
    tarfile.open(filepath, 'r:gz').extractall(MODEL_DIR)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2081, in extractall
    self.extract(tarinfo, path)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2118, in extract
    self._extract_member(tarinfo, os.path.join(path, tarinfo.name))
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2194, in _extract_member
    self.makefile(tarinfo, targetpath)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2235, in makefile
    copyfileobj(source, target)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 266, in copyfileobj
    shutil.copyfileobj(src, dst)
  File "/home/hyw/zp/anaconda2/lib/python2.7/shutil.py", line 63, in copyfileobj
    buf = fsrc.read(length)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 831, in read
    buf += self.fileobj.read(size - len(buf))
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 743, in read
    return self.readnormal(size)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 758, in readnormal
    return self.__read(size)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 748, in __read
    buf = self.fileobj.read(size)
  File "/home/hyw/zp/anaconda2/lib/python2.7/gzip.py", line 267, in read
    self._read(readsize)
  File "/home/hyw/zp/anaconda2/lib/python2.7/gzip.py", line 314, in _read
    self._read_eof()
  File "/home/hyw/zp/anaconda2/lib/python2.7/gzip.py", line 353, in _read_eof
    hex(self.crc)))
IOError: CRC check failed 0x49a91344 != 0x6adf6a7L
Traceback (most recent call last):
  File "score_mask.py", line 13, in <module>
    import tflib.inception_score
  File "/home/hyw/zp/tf/Pose-Guided-Person-Image-Generation-master/tflib/inception_score.py", line 101, in <module>
    _init_inception()
  File "/home/hyw/zp/tf/Pose-Guided-Person-Image-Generation-master/tflib/inception_score.py", line 73, in _init_inception
    tarfile.open(filepath, 'r:gz').extractall(MODEL_DIR)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2081, in extractall
    self.extract(tarinfo, path)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2118, in extract
    self._extract_member(tarinfo, os.path.join(path, tarinfo.name))
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2194, in _extract_member
    self.makefile(tarinfo, targetpath)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 2235, in makefile
    copyfileobj(source, target)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 266, in copyfileobj
    shutil.copyfileobj(src, dst)
  File "/home/hyw/zp/anaconda2/lib/python2.7/shutil.py", line 63, in copyfileobj
    buf = fsrc.read(length)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 831, in read
    buf += self.fileobj.read(size - len(buf))
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 743, in read
    return self.readnormal(size)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 758, in readnormal
    return self.__read(size)
  File "/home/hyw/zp/anaconda2/lib/python2.7/tarfile.py", line 748, in __read
    buf = self.fileobj.read(size)
  File "/home/hyw/zp/anaconda2/lib/python2.7/gzip.py", line 267, in read
    self._read(readsize)
  File "/home/hyw/zp/anaconda2/lib/python2.7/gzip.py", line 314, in _read
    self._read_eof()
  File "/home/hyw/zp/anaconda2/lib/python2.7/gzip.py", line 353, in _read_eof
    hex(self.crc)))
IOError: CRC check failed 0x49a91344 != 0x6adf6a7L

原因分析： 通过使用 vscode 中的 go to  defination 查看 import tflib.inception_score 中的 inception_score 模块中的函数，发现里面有个需要下载的文件链接，要下载  inception-2015-12-05.tgz，结合报错中出现了 tarflie 和 gzip 这些和解压压缩有关模块，所以初步怀疑，是文件不存在或者 下载的文件有缺陷造成的

到ubuntu 的 /tmp/imagenet 目录下查看 发现 inception-2015-12-05.tgz 这个文件只有 2.3M，而真实大小应该是 88.9M，所以怀疑该文件并未下载完整，导致的莫名错误

解决办法： 手动下载 inception-2015-12-05.tgz 文件，放到  /tmp/imagenet  目录下，问题解决

提醒一点： 报  IO错误    无非就是 文件  或者  路径目录  找不到

参考： python open打开文件失败 IOError: [Errno 2] No such file or directory: 'E://aaa.txt'