tf.keras.utils.text_dataset_from_directory

Generates a tf.data.Dataset from text files in a directory.

tf.keras.utils.text_dataset_from_directory(
    directory,
    labels='inferred',
    label_mode='int',
    class_names=None,
    batch_size=32,
    max_length=None,
    shuffle=True,
    seed=None,
    validation_split=None,
    subset=None,
    follow_links=False
)

If your directory structure is:

main_directory/
...class_a/
......a_text_1.txt
......a_text_2.txt
...class_b/
......b_text_1.txt
......b_text_2.txt

Then calling text_dataset_from_directory(main_directory, labels='inferred') will return a tf.data.Dataset that yields batches of texts from the subdirectories class_a and class_b, together with labels 0 and 1 (0 corresponding to class_a and 1 corresponding to class_b).

Only .txt files are supported at this time.

directory Directory where the data is located. If labels is "inferred", it should contain subdirectories, each containing text files for a class. Otherwise, the directory structure is ignored.
labels Either "inferred" (labels are generated from the directory structure), None (no labels), or a list/tuple of integer labels of the same size as the number of text files found in the directory. Labels should be sorted according to the alphanumeric order of the text file paths (obtained via os.walk(directory) in Python).
label_mode String describing the encoding of labels. Options are:

  • 'int': means that the labels are encoded as integers (e.g. for sparse_categorical_crossentropy loss).
  • 'categorical' means that the labels are encoded as a categorical vector (e.g. for categorical_crossentropy loss).
  • 'binary' means that the labels (there can be only 2) are encoded as float32 scalars with values 0 or 1 (e.g. for binary_crossentropy).
  • None (no labels).
class_names Only valid if "labels" is "inferred". This is the explicit list of class names (must match names of subdirectories). Used to control the order of the classes (otherwise alphanumerical order is used).
batch_size Size of the batches of data. Default: 32. If None, the data will not be batched (the dataset will yield individual samples).
max_length Maximum size of a text string. Texts longer than this will be truncated to max_length.
shuffle Whether to shuffle the data. Default: True. If set to False, sorts the data in alphanumeric order.
seed Optional random seed for shuffling and transformations.
validation_split Optional float between 0 and 1, fraction of data to reserve for validation.
subset Subset of the data to return. One of "training", "validation" or "both". Only used if validation_split is set. When subset="both", the utility returns a tuple of two datasets (the training and validation datasets respectively).
follow_links Whether to visits subdirectories pointed to by symlinks. Defaults to False.

A tf.data.Dataset object.

  • If label_mode is None, it yields string tensors of shape (batch_size,), containing the contents of a batch of text files.
  • Otherwise, it yields a tuple (texts, labels), where texts has shape (batch_size,) and labels follows the format described below.

Rules regarding labels format:

- if `label_mode` is `int`, the labels are an `int32` tensor of shape
  `(batch_size,)`.
- if `label_mode` is `binary`, the labels are a `float32` tensor of
  1s and 0s of shape `(batch_size, 1)`.
- if `label_mode` is `categorical`, the labels are a `float32` tensor
  of shape `(batch_size, num_classes)`, representing a one-hot
  encoding of the class index.