DNNCLassifier Tensorflow 上的 label_keys 类型错误

Question

我想将标签嵌入到 Tensorflow 中的 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)

File "embedding_model_probe.py", line 118, in <module>
    m.fit(input_fn=train_input_fn, steps=200)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/deprecation.py", line 281, in new_func
    return func(*args, **kwargs)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 430, in fit
    loss = self._train_model(input_fn=input_fn, hooks=hooks)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 927, in _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", line 1132, in _get_train_ops
    return self._call_model_fn(features, labels, model_fn_lib.ModeKeys.TRAIN)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/estimator.py", line 1103, in _call_model_fn
    model_fn_results = self._model_fn(features, labels, **kwargs)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn.py", line 180, in _dnn_model_fn
    logits=logits)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py", line 1004, in create_model_fn_ops
    labels = self._transform_labels(mode=mode, labels=labels)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py", line 1033, in _transform_labels
    "label_ids": table.lookup(labels_tensor),
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/lookup/lookup_ops.py", line 179, in lookup
    (self._key_dtype, keys.dtype))
  TypeError: Signature mismatch. Keys must be dtype 

< dtype: 'string'>, got < dtype: 'int64'>

另一方面，如果我将

label_key_values

列设置为

numpy.array

，我将收到以下错误：

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

Traceback (most recent call last):
  File "embedding_model_probe.py", line 116, in <module>
    label_keys=label_keys_values)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/dnn.py", line 337, in __init__
    label_keys=label_keys),
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py", line 331, in multi_class_head
    label_keys=label_keys)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/learn/python/learn/estimators/head.py", line 986, in __init__
    if label_keys and len(label_keys) != n_classes:
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Answer 1

我找到了解决方案。正如官方文档所说here：

如果用户在构造函数中指定 label_keys，则标签必须是 label_keys 词汇表中的字符串。

在我的例子中，我将训练和测试集中的标签列转换为一个单热向量（整数值），并且

label_keys_values

数组中的值与它们不匹配。