Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings


View source on GitHub

Converts a tf.data.Dataset to an iterable of NumPy arrays.

    dataset, graph=None

as_numpy converts a possibly nested structure of tf.data.Datasets and tf.Tensors to iterables of NumPy arrays and NumPy arrays, respectively.

Note that because TensorFlow has support for ragged tensors and NumPy has no equivalent representation, tf.RaggedTensors are left as-is for the user to deal with them (e.g. using to_list()). In TF 1 (i.e. graph mode), tf.RaggedTensors are returned as tf.ragged.RaggedTensorValues.


ds = tfds.load(name="mnist", split="train")
ds_numpy = tfds.as_numpy(ds)  # Convert `tf.data.Dataset` to Python generator
for ex in ds_numpy:
  # `{'image': np.array(shape=(28, 28, 1)), 'labels': np.array(shape=())}`



A structure matching dataset where tf.data.Datasets are converted to generators of NumPy arrays and tf.Tensors are converted to NumPy arrays.