How to load sparse data with TensorFlow?

Question

There is a small snippet about loading sparse data but I have no idea how to use it.

SparseTensors don\'t play well with queues. If you use SparseTensors

Answer 1

You can use weighted_categorical_column to parse index and value, eg.

categorical_column = tf.feature_column.categorical_column_with_identity(
            key='index', num_buckets=your_feature_dim)
sparse_columns = tf.feature_column.weighted_categorical_column(
    categorical_column=categorical_column, weight_feature_key='value')

then feed sparse_columns to linear model estimator, before feed to DNN, please use embedding, eg.

dense_columns = tf.feature_column.embedding_column(sparse_columns, your_embedding_dim)

then feed dense_columns to your DNN estimator