Table of Contents
Contents

FiftyOne Dataset Zoo

FiftyOne provides a Dataset Zoo that contains a collection of common datasets that you can download and load into FiftyOne via a few simple commands.

Note

For some datasets, FiftyOne’s Dataset Zoo uses the TorchVision Datasets or TensorFlow Datasets, depending on which ML library you have installed.

If you do not have the proper packages installed when attempting to download a zoo dataset, you will receive an error message that will help you resolve the issue. See customizing your ML backend for more information about configuring the backend behavior of the Dataset Zoo.

Available datasets

The Dataset Zoo contains dozens of datasets that you can load into FiftyOne with a few simple commands. Click the link below to see all of the datasets in the zoo!

Explore the datasets in the zoo

API reference

The Dataset Zoo can be accessed via Python library and the CLI. Consult the API reference below to see how to download, load, and manage zoo datasets.

Check out the API reference

Basic recipe

Methods for working with the Dataset Zoo are conveniently exposed via the Python library and the CLI. The basic recipe for loading a zoo dataset and visualizing it in the App is shown below.

Python

CLI

Use load_zoo_dataset() to load a zoo dataset into a FiftyOne dataset.

For example, the code sample below loads the validation split of COCO-2017 from the zoo and visualizes it in the FiftyOne App:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
import fiftyone as fo
import fiftyone.zoo as foz

# List available zoo datasets
print(foz.list_zoo_datasets())

#
# Load the COCO-2017 validation split into a FiftyOne dataset
#
# This will download the dataset from the web, if necessary
#
dataset = foz.load_zoo_dataset("coco-2017", split="validation")

# Give the dataset a new name, and make it persistent so that you can
# work with it in future sessions
dataset.name = "coco-2017-validation-example"
dataset.persistent = True

# Visualize the in the App
session = fo.launch_app(dataset)

Use fiftyone zoo datasets load to load a zoo dataset into a FiftyOne dataset.

For example, the code sample below loads the validation split of COCO-2017 from the zoo and visualizes it in the FiftyOne App:

#
# Load the COCO-2017 validation split into a FiftyOne dataset called
# `coco-2017-validation-example`
#
# This will download the dataset from the web, if necessary
#
fiftyone zoo datasets load coco-2017 --split validation \
    --dataset-name coco-2017-validation-example

# Visualize the dataset in the App
fiftyone app launch coco-2017-validation-example
Dataset Zoo