Save the date! Google I/O returns May 18-20 Register now

Returns an index lookup table based on the given dataset.

This operation constructs a lookup table based on the given dataset of keys.

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:

ds = x: tf.strings.as_string(x * 2))
table =
                                    ds, key_dtype=dtypes.int64)
table.lookup(tf.constant(['0', '2', '4'], dtype=tf.string)).numpy()
array([0, 1, 2])

dataset A dataset of keys.
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

  • num_oov_buckets is negative
  • vocab_size is not greater than zero
  • The key_dtype is not integer or string