张量板:文件系统方案 gs 未实现



我无法tensorboard连接到我的谷歌云平台,因为我面临以下错误:

我正在运行的命令:

gcloud auth application-default logintensorboard --logdir=gs://mybucket_which_contains_train_and_eval_directories

堆栈跟踪:

Exception in thread Reloader:
Traceback (most recent call last):
File "c:pythonpython35libthreading.py", line 914, in _bootstrap_inner
self.run()
File "c:pythonpython35libthreading.py", line 862, in run
self._target(*self._args, **self._kwargs)
File "c:pythonpython35libsite-packagestensorboardbackendapplication.py", line 327, in _reload_forever
reload_multiplexer(multiplexer, path_to_run)
File "c:pythonpython35libsite-packagestensorboardbackendapplication.py", line 299, in reload_multiplexer
multiplexer.AddRunsFromDirectory(path, name)
File "c:pythonpython35libsite-packagestensorboardbackendevent_processingevent_multiplexer.py", line 175, in AddRunsFromDirectory
for subdir in GetLogdirSubdirectories(path):
File "c:pythonpython35libsite-packagestensorboardbackendevent_processingevent_multiplexer.py", line 439, in GetLogdirSubdirectories
if tf.gfile.Exists(path) and not tf.gfile.IsDirectory(path):
File "c:pythonpython35libsite-packagestensorflowpythonlibiofile_io.py", line 252, in file_exists
pywrap_tensorflow.FileExists(compat.as_bytes(filename), status)
File "c:pythonpython35libcontextlib.py", line 66, in __exit__
next(self.gen)
File "c:pythonpython35libsite-packagestensorflowpythonframeworkerrors_impl.py", line 466, in raise_exception_on_not_ok_status
pywrap_tensorflow.TF_GetCode(status))
tensorflow.python.framework.errors_impl.UnimplementedError: File system scheme gs not implemented`

操作系统平台和发行版(例如,Linux Ubuntu 16.04(:Windows 10

TensorFlow 安装自(源或二进制(:pip3 install tensorflow

TensorFlow版本(使用下面的命令(:1.3.0

张量板版本:0.1.4

蟒蛇版本:3.5.2

似乎 tensorflow 还不支持 Windowsgs文件系统。遵循以下 git 问题:

  • https://github.com/tensorflow/tensorflow/issues/11280
  • https://github.com/tensorflow/tensorboard/issues/429

最新更新