ML Community Day is November 9! Join us for updates from TensorFlow, JAX, and more Learn more

tf.data.experimental.table_from_dataset

Returns a lookup table based on the given dataset.

This operation constructs a lookup table based on the given dataset of pairs of (key, value).

Any lookup of an out-of-vocabulary token will return a bucket ID based on its hash if num_oov_buckets is greater than zero. Otherwise it is assigned the default_value. The bucket ID range is [vocabulary size, vocabulary size + num_oov_buckets - 1].

Sample Usages:

keys = tf.data.Dataset.range(100)
values = tf.data.Dataset.range(100).map(
    lambda x: tf.strings.as_string(x * 2))
ds = tf.data.Dataset.zip((keys, values))
table = tf.data.experimental.table_from_dataset(
                              ds, default_value='n/a', key_dtype=tf.int64)
table.lookup(tf.constant([0, 1, 2], dtype=tf.int64)).numpy()
array([b'0', b'2', b'4'], dtype=object)

dataset A dataset containing (key, value) pairs.
num_oov_buckets The number of out-of-vocabulary buckets.
vocab_size Number of the elements in the vocabulary, if known.
default_value The value to use for out-of-vocabulary feature values. Defaults to -1.
hasher_spec A HasherSpec to specify the hash function to use for assignation of out-of-vocabulary buckets.
key_dtype The key data type.
name A name for this op (optional).

The lookup table based on the given dataset.

ValueError If

  • dataset does not contain pairs
  • The 2nd item in the dataset pairs has a dtype which is incompatible with default_value
  • num_oov_buckets is negative
  • vocab_size is not greater than zero
  • The key_dtype is not integer or string