TensorFlow Deep and Wide Demo 问题

Question

我尝试直接从tensorflow的Deep and Wide demo repo运行代码：

urllib 存在一个直接问题，可以通过使用 urllib.request 轻松解决。之后代码仍然无法运行，我收到以下错误：

m.fit(input_fn=lambda: input_fn(df_train), steps=FLAGS.train_steps)
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/Users/USER/tensorflow/lib/python3.5/site-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 182, in fit
monitors=monitors)
File "/Users/USER/tensorflow/lib/python3.5/site-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 449, in _train_model
train_op, loss_op = self._get_train_ops(features, targets)
File "/Users/USER/tensorflow/lib/python3.5/site-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 156, in _get_train_ops
logits = self._logits(features, is_training=True)
File "/Users/USER/tensorflow/lib/python3.5/site-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 294, in _logits
if self._get_linear_feature_columns() and self._get_dnn_feature_columns():
File "/Users/USER/tensorflow/lib/python3.5/site-packages/tensorflow/contrib/learn/python/learn/estimators/dnn_linear_combined.py", line 216, in _get_dnn_feature_columns
self._dnn_feature_columns)) if self._dnn_feature_columns else None
TypeError: unorderable types: str() < _SparseColumnKeys()

我无法找到此问题的根源。似乎没有其他人遇到此问题。 TensorFlow 安装在 python 3.5 上的 virtualenv (tensorflow) 中。

Answer 1

urllib 适用于 2.7，请尝试在 2.7 而不是 3.5 中运行它。

Answer 2

我遇到了同样的问题。这样就解决了。

我必须在 Python 2.7 下安装 Tensorflow。以下是使用 Conda 在虚拟环境中执行此操作的方法：

# Python 2.7
$ conda create -n tensorflow python=2.7

$ source activate tensorflow
(tensorflow)$  # Your prompt should change

# Linux/Mac OS X, Python 2.7/3.4/3.5, CPU only:
(tensorflow)$ conda install -c conda-forge tensorflow

查看：https://www.tensorflow.org/versions/r0.10/get_started/os_setup.html#using-conda