There are a few different ways to load a dataset in TensorFlow.
Option 1: Load a dataset from a file
To load a dataset from a file, you can use the tf.data.TextLineDataset class. This class can be used to load text files, CSV files, and other types of files.
For example, to load a CSV file, you can use the following code:
import tensorflow as tf
dataset = tf.data.TextLineDataset("data.csv")
This will create a dataset that contains the lines from the file data.csv.
Option 2: Load a dataset from a URL
To load a dataset from a URL, you can use the tf.data.WebDataset class. This class can be used to load datasets from the internet.
For example, to load a dataset from the MNIST dataset, you can use the following code:
import tensorflow as tf
dataset = tf.data.WebDataset("https://storage.googleapis.com/tensorflow/tf-datasets/mnist.tfrecord")
This will create a dataset that contains the MNIST dataset.
Option 3: Load a dataset from a TensorFlow Datasets collection
TensorFlow Datasets is a collection of pre-built datasets that can be used with TensorFlow. To load a dataset from TensorFlow Datasets, you can use the tfds.load function.
For example, to load the MNIST dataset from TensorFlow Datasets, you can use the following code:
import tensorflow as tf
import tensorflow_datasets as tfds
dataset = tfds.load("mnist")
This will create a dataset that contains the MNIST dataset.
Once you have loaded a dataset, you can iterate over it to access the data. To iterate over a dataset, you can use the for loop.
For example, to iterate over the MNIST dataset, you can use the following code:
import tensorflow as tf
import tensorflow_datasets as tfds
dataset = tfds.load("mnist")
for image, label in dataset:
print(image.shape, label)
This will print the shape of the image and the label for each image in the dataset.
You can also use the tf.data.Dataset class to manipulate the data in a dataset. For example, you can use the map method to apply a function to each element in the dataset.
For example, to convert the images in the MNIST dataset to grayscale, you can use the following code:
import tensorflow as tf
import tensorflow_datasets as tfds
dataset = tfds.load("mnist")
dataset = dataset.map(lambda image: tf.image.rgb_to_grayscale(image))
for image, label in dataset:
print(image.shape, label)
This will print the shape of the grayscale image and the label for each image in the dataset.
