谷歌开源机器学习数据集，可在 TensorFlow 直接调用

# Install: pip install tensorflow-datasetsimport tensorflow_datasets as tfdsmnist_data = tfds.load("mnist")mnist_train, mnist_test = mnist_data["train"], mnist_data["test"]assert isinstance(mnist_train, tf.data.Dataset)

import tensorflow_datasets as tfds
# Fetch the dataset directlymnist = tfds.image.MNIST()# or by string namemnist = tfds.builder('mnist')
# Describe the dataset with DatasetInfoassert mnist.info.features['image'].shape == (28, 28, 1)assert mnist.info.features['label'].num_classes == 10assert mnist.info.splits['train'].num_examples == 60000
# Download the data, prepare it, and write it to diskmnist.download_and_prepare()
# Load data from disk as tf.data.Datasetsdatasets = mnist.as_dataset()train_dataset, test_dataset = datasets['train'], datasets['test']assert isinstance(train_dataset, tf.data.Dataset)
# And convert the Dataset to NumPy arrays if you'd likefor example in tfds.as_numpy(train_dataset):  image, label = example['image'], example['label']  assert isinstance(image, np.array)

import tensorflow_datasets as tfds
datasets = tfds.load("mnist")train_dataset, test_dataset = datasets["train"], datasets["test"]assert isinstance(train_dataset, tf.data.Dataset)

# See the built-in configsconfigs = tfds.text.IMDBReviews.builder_configsassert "bytes" in configs
# Address a built-in config with tfds.builderimdb = tfds.builder("imdb_reviews/bytes")# or when constructing the builder directlyimdb = tfds.text.IMDBReviews(config="bytes")# or use your own custom configurationmy_encoder = tfds.features.text.ByteTextEncoder(additional_tokens=['hello'])my_config = tfds.text.IMDBReviewsConfig(    name="my_config",    version="1.0.0",    text_encoder_config=tfds.features.text.TextEncoderConfig(encoder=my_encoder),)imdb = tfds.text.IMDBReviews(config=my_config)

imdb = tfds.builder("imdb_reviews/subwords8k")
# Get the TextEncoder from DatasetInfoencoder = imdb.info.features["text"].encoderassert isinstance(encoder, tfds.features.text.SubwordTextEncoder)
# Encode, decodeids = encoder.encode("Hello world")assert encoder.decode(ids) == "Hello world"
# Get the vocabulary sizevocab_size = encoder.vocab_size