tf.keras.layers.TextVectorization

Text vectorization layer.

Inherits From: PreprocessingLayer, Layer, Module

Used in the notebooks

Used in the guide Used in the tutorials

This layer has basic options for managing text in a Keras model. It transforms a batch of strings (one example = one string) into either a list of token indices (one example = 1D tensor of integer token indices) or a dense representation (one example = 1D tensor of float values representing data about the example's tokens).

If desired, the user can call this layer's adapt() method on a dataset. When this layer is adapted, it will analyze the dataset, determine the frequency of individual string values, and create a 'vocabulary' from them. This vocabulary can have unlimited size or be capped, depending on the configuration options for this layer; if there are more unique values in the input than the maximum vocabulary size, the most frequent terms will be used to create the vocabulary.

The processing of each example contains the following steps:

  1. Standardize each example (usually lowercasing + punctuation stripping)
  2. Split each example into substrings (usually words)
  3. Recombine substrings into tokens (usually ngrams)
  4. Index tokens (associate a unique int value with each token)
  5. Transform each example using this index, either into a vector of ints or a dense float vector.

Some notes on passing callables to customize splitting and normalization for this layer:

  1. Any callable can be passed to this Layer, but if you want to serialize this object you should only pass functions that are registered Keras serializables (see tf.keras.utils.register_keras_serializable for more details).
  2. When using a custom callable for standardize, the data received by the callable will be exactly as passed to this layer. The callable should return a tensor of the same shape as the input.
  3. When using a custom callable for split, the data received by the callable will have the 1st dimension squeezed out - instead of [["string to split"], ["another string to split"]], the Callable will see ["string to split", "another string to split"]. The callable should return a Tensor with the first dimension containing the split tokens - in this example, we should see something like [["string", "to", "split"], ["another", "string", "to", "split"]]. This makes the callable site natively compatible with tf.strings.split().

max_tokens The maximum size of the vocabulary for this layer. If None, there is no cap on the size of the vocabulary. Note that this vocabulary contains 1 OOV token, so the effective number of tokens is (max_tokens - 1 - (1 if output_mode == "int" else 0)).
standardize Optional specification for standardization to apply to the input text. Values can be None (no standardization), "lower_and_strip_punctuation" (lowercase and remove punctuation) or a Callable. Default is "lower_and_strip_punctuation".
split Optional specification for splitting the input text. Values can be None (no splitting), "whitespace" (split on ASCII whitespace), or a Callable. The default is "whitespace".
ngrams Optional specification for ngrams to create from the possibly-split input text. Values can be None, an integer or tuple of integers; passing an integer will create ngrams up to that integer, and passing a tuple of integers will create ngrams for the specified values in the tuple. Passing None means that no ngrams will be created.
output_mode Optional specification for the output of the layer. Values can be "int", "multi_hot", "count" or "tf_idf", configuring the layer as follows:

  • "int": Outputs integer indices, one integer index per split string token. When output_mode == "int", 0 is reserved for masked locations; this reduces the vocab size to max_tokens - 2 instead of max_tokens - 1.
  • "multi_hot": Outputs a single int array per batch, of either vocab_size or max_tokens size, containing 1s in all elements where the token mapped to that index exists at least once in the batch item.
  • "count": Like "multi_hot", but the int array contains a count of the number of times the token at that index appeared in the batch item.
  • "tf_idf": Like "multi_hot", but the TF-IDF algorithm is applied to find the value in each token slot. For "int" output, any shape of input and output is supported. For all other output modes, currently only rank 1 inputs (and rank 2 outputs after splitting) are supported.
output_sequence_length Only valid in INT mode. If set, the output will have its time dimension padded or truncated to exactly output_sequence_length values, resulting in a tensor of shape (batch_size, output_sequence_length) regardless of how many tokens resulted from the splitting step. Defaults to None.
pad_to_max_tokens Only valid in "multi_hot", "count", and "tf_idf" modes. If True, the output will have its feature axis padded to max_tokens even if the number of unique tokens in the vocabulary is less than max_tokens, resulting in a tensor of shape (batch_size, max_tokens) regardless of vocabulary size. Defaults to False.
vocabulary Optional. Either an array of strings or a string path to a text file. If passing an array, can pass a tuple, list, 1D numpy array, or 1D tensor containing the string vocbulary terms. If passing a file path, the file should contain one line per term in the vocabulary. If this argument is set, there is no need to adapt the layer.

Example:

This example instantiates a TextVectorization layer that lowercases text, splits on whitespace, strips punctuation, and outputs integer vocab indices.

text_dataset = tf.data.Dataset.from_tensor_slices(["foo", "bar", "baz"])