CODEWITHSUNDEEP
tensorflowmachine-learning
Genarito
Genarito

Reputation: 3453

Items of feature_columns must be a _FeatureColumn Given: _VocabularyListCategoricalColumn

I'm trying to make my_feature_columns list from this dataset. I'm reading it with this code:

train = pd.read_csv(train_path, names=CSV_COLUMN_NAMES, header=0)
train = train.drop(['Name', 'SibSp', 'Parch', 'Ticket', 'Fare', 'Cabin', 'Embarked'], axis=1)
train.fillna(0, inplace=True)

I just want to use "Pclass", "Sex" and "Age". Since Sex only have 2 possible values I write this code:

my_feature_columns = [
    tf.feature_column.numeric_column(key='Pclass'),
    tf.feature_column.categorical_column_with_vocabulary_list(key="Sex", vocabulary_list=["male", "female"]),
    tf.feature_column.numeric_column(key='Age'),
]

Then I instanciates the Classifier:

classifier = tf.estimator.DNNClassifier(
        feature_columns=my_feature_columns,
        hidden_units=[10, 10],
        n_classes=2)

But I'm getting

ValueError: Items of feature_columns must be a _DenseColumn. You can wrap a categorical column with an embedding_column or indicator_column. Given: _VocabularyListCategoricalColumn(key='Sex', vocabulary_list=('male', 'female'), dtype=tf.string, default_value=0, num_oov_buckets=0)

I'm totally new with TensorFlow and I have no idea what's going on. I'll apreciate all kind of help, thanks in advance and sorry about my English.

Upvotes: 14

Views: 8522

Answers (2)

Hamid
Hamid

Reputation: 231

There is another way According to This Source Code:

So:

categorical_column = tf.feature_column.categorical_column_with_vocabulary_list(key="Sex", vocabulary_list=["male", "female"], default_value=0)

my_feature_columns = [
    tf.feature_column.numeric_column(key='Pclass'),
    embedded_group_column = tf.feature_column.embedding_column(categorical_column,dimension=number_of_categories)
    tf.feature_column.numeric_column(key='Age')
]

Good Luck!

Upvotes: 2

Genarito
Genarito

Reputation: 3453

Reading again the documentation I realised that

Indicator columns and embedding columns never work on features directly

So I corrected my code:

categorical_column = tf.feature_column.categorical_column_with_vocabulary_list(key="Sex", vocabulary_list=["male", "female"], default_value=0)

my_feature_columns = [
    tf.feature_column.numeric_column(key='Pclass'),
    tf.feature_column.indicator_column(categorical_column),
    tf.feature_column.numeric_column(key='Age')
]

And works like a charm!

Upvotes: 25

Related Questions