带有 tf-idf 稀疏矩阵的 Tensorflow DNN

Question

正在尝试实现用于文本分类的 tesorflow DNN。

tf-idf 稀疏 IV:

X_train_sam:
<31819x3122 sparse matrix of type '<class 'numpy.float64'>'with 610128 stored elements in Compressed Sparse Row format>

标记为 DV：

y_train_sam.values:array(['mexican', 'mexican', 'italian', ..., 'chinese', 'italian','italian'], dtype=object)

使用以下片段将稀疏转换为张量：

def convert_sparse_matrix_to_sparse_tensor(X):
    coo = X.tocoo()
    indices = np.mat([coo.row, coo.col]).transpose()
    return tf.SparseTensorValue(indices, coo.data, coo.shape)

 X_train_sam = convert_sparse_matrix_to_sparse_tensor(X_train_sam)

正在为建模准备数据

def train_input_fn(features, labels, batch_size):
    dataset = tf.data.Dataset.from_tensors((features, labels))
    dataset = dataset.shuffle(1000).repeat().batch(batch_size)
    return dataset.make_one_shot_iterator().get_next()

inp = train_input_fn(X_train_sam,y_train_sam.values,batch_size=1000)

应用 DNN 分类器

classifier = tf.estimator.DNNClassifier(
    feature_columns=[float]*X_train_sam.dense_shape[1],
    hidden_units=[10, 10],
    n_classes=len(y_train_sam.unique()))

classifier.train(input_fn=lambda:inp)

出现以下错误：

ValueError: features should be a dictionary of `Tensor`s. Given type: <class 'tensorflow.python.framework.sparse_tensor.SparseTensorValue'>

请指点，我是ML和tensorflow的新手。

Answer 1

如果在你的代码中这一行

classifier.train(input_fn=lambda:inp)

lambda:inp 应该是字典还是你的意思是匿名函数？来自

的文档

https://www.tensorflow.org/api_docs/python/tf/estimator/DNNClassifier

input_fn: Input function returning a tuple of: features - Tensor or dictionary of string feature name to Tensor. labels - Tensor or dictionary of Tensor with labels.

所以你需要一个 returns 元组而不是单个值的函数...

带有 tf-idf 稀疏矩阵的 Tensorflow DNN

Tensorflow DNN with tf-idf sparse matrix

machine-learning

tensorflow

tensorflow-datasets