Coursera

TFDS Hello World

#@title Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

In this notebook we will take a look at the simple Hello World scenario of TensorFlow Datasets (TFDS). We’ll use TFDS to perform the extract, transform, and load processes for the MNIST dataset.

Setup

We’ll start by importing TensorFlow, TensorFlow Datasets, and Matplotlib.

try:
    %tensorflow_version 2.x
except:
    pass

Colab only includes TensorFlow 2.x; %tensorflow_version has no effect.

import matplotlib.pyplot as plt

import tensorflow as tf
import tensorflow_datasets as tfds

print("\u2022 Using TensorFlow Version:", tf.__version__)

• Using TensorFlow Version: 2.14.0

Extract - Transform - Load (ETL)

Now we’ll run the ETL code. First, to perform the Extract process we use tfts.load. This handles everything from downloading the raw data to parsing and splitting it, giving us a dataset. Next, we perform the Transform process. In this simple example, our transform process will just consist of shuffling the dataset. Finally, we Load one record by using the take(1) method. In this case, each record consists of an image and its corresponding label. After loading the record we proceed to plot the image and print its corresponding label.

# EXTRACT
dataset = tfds.load(name="mnist", split="train")
# TRANSFORM
dataset.shuffle(100)

Downloading and preparing dataset 11.06 MiB (download: 11.06 MiB, generated: 21.00 MiB, total: 32.06 MiB) to /root/tensorflow_datasets/mnist/3.0.1...



Dl Completed...:   0%|          | 0/5 [00:00<?, ? file/s]


Dataset mnist downloaded and prepared to /root/tensorflow_datasets/mnist/3.0.1. Subsequent calls will reuse this data.





<_ShuffleDataset element_spec={'image': TensorSpec(shape=(28, 28, 1), dtype=tf.uint8, name=None), 'label': TensorSpec(shape=(), dtype=tf.int64, name=None)}>

# LOAD
for data in dataset.take(1):
    image = data["image"].numpy().squeeze()
    label = data["label"].numpy()

    print("Label: {}".format(label))
    plt.imshow(image, cmap=plt.cm.binary)
    plt.show()

Label: 4