tensorflow学习笔记:运行tensorboard遇到的错误

在virtualenv环境下执行tensorboard时遇到错误：

错误：

(tensorflow) fcx@fcx:~/share/test/deeplearning$ tensorboard --logdir = './save_model_data_onehide/train_logs/'

Traceback (most recent call last):
  File "/home/fcx/share/test/deeplearning/tensorflow/bin/tensorboard", line 161, in <module>
    Main()
  File "/home/fcx/share/test/deeplearning/tensorflow/bin/tensorboard", line 111, in Main
    module_space = FindModuleSpace()
  File "/home/fcx/share/test/deeplearning/tensorflow/bin/tensorboard", line 78, in FindModuleSpace
    for mod in site.getsitepackages():
AttributeError: 'module' object has no attribute 'getsitepackages'

解决：

打开在virtualenv环境目录下xxx/bin/tensorboard文件，将69行处的

for mod in site.getsitepackages():

替换为如下内容：

if hasattr(site, 'getsitepackages'):
        # normal execution
        sitepackages = site.getsitepackages()
    else:
        # workaround for virtualenv
        from distutils.sysconfig import get_python_lib
        sitepackages = [get_python_lib()]

    #for mod in site.getsitepackages():
    for mod in sitepackages:
      module_space = mod + '/tensorboard/tensorboard' + '.runfiles'
      if os.path.isdir(module_space):
        return module_space

接下来又遇到这个错误，在google上查了好久都没找到答案

(tensorflow) fcx@fcx:~/share/test/deeplearning$ tensorboard --logdir = './save_model_data_onehide/train_logs/'
F tensorflow/core/platform/cpu_feature_guard.cc:35] The TensorFlow library was compiled to use SSE4.1 instructions, but these aren't available on your machine.
已放弃 (核心已转储)

只好退出virtualenv环境，然后upgrade下tensorflow

sudo pip install --upgrade tensorflow

升级完tensorflow后在normal环境下运行tensorboard就不报错了。

但是打开浏览器查看提示：No dashboards are active for the current data set.没有加载到数据。郁闷。。。

尝试了下，切换到存放数据的目录然后运行：

tensorboard --logdir = ./ 就OK了

特此记录下。