DNNCLassifier Tensorflow上的label_keys类型错误

我想在Tensorflow中将标签嵌入到DNNClassifier模型中 . 与文档示例here不同，我收到以下错误消息：

label_keys_values = ["satan", "ipsweep", "nmap", "portsweep"]  
m = tf.contrib.learn.DNNClassifier(model_dir=model_dir,
                                  feature_columns=deep_columns,
                                  n_classes=4,
                                  hidden_units=[12, 4],
                                  label_keys=label_keys_values)
m.fit(input_fn=train_input_fn, steps=200)

文件“embedding_model_probe.py”，第118行，在m.fit（input_fn = train_input_fn，steps = 200）文件“/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/deprecation.py “，第281行，在new_func中返回func（* args，** kwargs）文件”/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py“ ，第430行，在fit loss = self._train_model（input_fn = input_fn，hooks = hooks）文件“/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator .py“，第927行，_train_model model_fn_ops = self._get_train_ops（features，labels）File”/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator . py“，第1132行，在_get_train_ops中返回self._call_model_fn（features，labels，model_fn_lib.ModeKeys.TRAIN）文件”/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/ estimators / estimator.py“，第1103行，在_call_model_fn中model_fn_results = self._model_fn（功能，标签，* * kwargs）文件“/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn.py”，第180行，在_dnn_model_fn中logits = logits）文件“/ usr /local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py“，第1004行，在create_model_fn_ops labels = self._transform_labels（mode = mode，labels = labels）文件“/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py”，第1033行，在_transform_labels“label_ids”中：table.lookup（labels_tensor），File “/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/lookup/lookup_ops.py”，第179行，在查找中（self._key_dtype，keys.dtype））TypeError：签名不匹配 . 键必须是dtype

<dtype：'string'>，得到<dtype：'int64'>

另一方面，如果我将 label_key_values 列设为 numpy.array ，我将收到以下错误：

label_keys_values = np.array(["satan", "ipsweep", "nmap", "portsweep"], dtype='string')

回溯（最近一次调用最后一次）：文件“embedding_model_probe.py”，第116行，在label_keys = label_keys_values中）文件“/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn /estimators/dnn.py“，第337行，在init label_keys = label_keys中），文件”/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py “，第331行，在multi_class_head label_keys = label_keys中）文件”/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py“，第986行，在init中if label_keys和len（label_keys）！= n_classes：ValueError：具有多个元素的数组的真值是不明确的 . 使用a.any（）或a.all（）