fiftyone.zoo.datasets¶

Module contents¶

The FiftyOne Dataset Zoo.

This package defines a collection of open source datasets made available for download via FiftyOne.

Copyright 2017-2024, Voxel51, Inc.
voxel51.com

Functions:

`list_zoo_datasets`([tags, source])	Returns the list of available datasets in the FiftyOne Dataset Zoo.
`list_downloaded_zoo_datasets`([base_dir])	Returns information about the zoo datasets that have been downloaded.
`download_zoo_dataset`(name[, split, splits, …])	Downloads the dataset of the given name from the FiftyOne Dataset Zoo.
`load_zoo_dataset`(name[, split, splits, …])	Loads the dataset of the given name from the FiftyOne Dataset Zoo as a `fiftyone.core.dataset.Dataset`.
`find_zoo_dataset`(name[, split])	Returns the directory containing the given zoo dataset.
`load_zoo_dataset_info`(name[, dataset_dir])	Loads the `ZooDatasetInfo` for the specified zoo dataset.
`get_zoo_dataset`(name, **kwargs)	Returns the `ZooDataset` instance for the dataset with the given name.
`delete_zoo_dataset`(name[, split])	Deletes the zoo dataset from local disk, if necessary.

Classes:

`ZooDatasetInfo`(zoo_dataset, dataset_type, …)	Class containing info about a dataset in the FiftyOne Dataset Zoo.
`ZooDatasetSplitInfo`(split, num_samples)	Class containing info about a split of a dataset in the FiftyOne Dataset Zoo.
`ZooDataset`()	Base class for datasets made available in the FiftyOne Dataset Zoo.
`DeprecatedZooDataset`()	Class representing a zoo dataset that no longer exists in the FiftyOne Dataset Zoo.

fiftyone.zoo.datasets.list_zoo_datasets(tags=None, source=None)¶

Returns the list of available datasets in the FiftyOne Dataset Zoo.

Example usage:

import fiftyone as fo
import fiftyone.zoo as foz

#
# List all zoo datasets
#

names = foz.list_zoo_datasets()
print(names)

#
# List all zoo datasets with (both of) the specified tags
#

names = foz.list_zoo_datasets(tags=["image", "detection"])
print(names)

#
# List all zoo datasets available via the given source
#

names = foz.list_zoo_datasets(source="torch")
print(names)

Parameters

tags (None) – only include datasets that have the specified tag or list of tags
source (None) – only include datasets available via the given source or list of sources

Returns

a sorted list of dataset names

fiftyone.zoo.datasets.list_downloaded_zoo_datasets(base_dir=None)¶

Returns information about the zoo datasets that have been downloaded.

Parameters: base_dir (None) – the base directory to search for downloaded datasets. By default, fo.config.dataset_zoo_dir is used
Returns: a dict mapping dataset names to (dataset dir, ZooDatasetInfo) tuples

fiftyone.zoo.datasets.download_zoo_dataset(name, split=None, splits=None, dataset_dir=None, overwrite=False, cleanup=True, **kwargs)¶

Downloads the dataset of the given name from the FiftyOne Dataset Zoo.

Any dataset splits that already exist in the specified directory are not re-downloaded, unless overwrite == True is specified.

Parameters

name – the name of the zoo dataset to download. Call list_zoo_datasets() to see the available datasets
split (None) – ("train", "validation", "test"). If neither split nor splits are provided, all available splits are downloaded. Consult the documentation for the ZooDataset you specified to see the supported splits
splits (None) – a list of splits to download, if applicable. Typical values are ("train", "validation", "test"). If neither split nor splits are provided, all available splits are downloaded. Consult the documentation for the ZooDataset you specified to see the supported splits
dataset_dir (None) – the directory into which to download the dataset. By default, it is downloaded to a subdirectory of fiftyone.config.dataset_zoo_dir
overwrite (False) – whether to overwrite any existing files
cleanup (True) – whether to cleanup any temporary files generated during download
**kwargs – optional arguments for the ZooDataset constructor

Returns

tuple of

info: the ZooDatasetInfo for the dataset
dataset_dir: the directory containing the dataset

fiftyone.zoo.datasets.load_zoo_dataset(name, split=None, splits=None, label_field=None, dataset_name=None, dataset_dir=None, download_if_necessary=True, drop_existing_dataset=False, persistent=False, overwrite=False, cleanup=True, progress=None, **kwargs)¶

Loads the dataset of the given name from the FiftyOne Dataset Zoo as a fiftyone.core.dataset.Dataset.

By default, the dataset will be downloaded if it does not already exist in the specified directory.

If you do not specify a custom dataset_name and you have previously loaded the same zoo dataset and split(s) into FiftyOne, the existing fiftyone.core.dataset.Dataset will be returned.

Parameters

name – the name of the zoo dataset to load. Call list_zoo_datasets() to see the available datasets
split (None) – ("train", "validation", "test"). If neither split nor splits are provided, all available splits are loaded. Consult the documentation for the ZooDataset you specified to see the supported splits
splits (None) – a list of splits to load, if applicable. Typical values are ("train", "validation", "test"). If neither split nor splits are provided, all available splits are loaded. Consult the documentation for the ZooDataset you specified to see the supported splits
label_field (None) – the label field (or prefix, if the dataset contains multiple label fields) in which to store the dataset’s labels. By default, this is "ground_truth" if the dataset contains a single label field. If the dataset contains multiple label fields and this value is not provided, the labels will be stored under dataset-specific field names
dataset_name (None) – an optional name to give the returned fiftyone.core.dataset.Dataset. By default, a name will be constructed based on the dataset and split(s) you are loading
dataset_dir (None) – the directory in which the dataset is stored or will be downloaded. By default, the dataset will be located in fiftyone.config.dataset_zoo_dir
download_if_necessary (True) – whether to download the dataset if it is not found in the specified dataset directory
drop_existing_dataset (False) – whether to drop an existing dataset with the same name if it exists
persistent (False) – whether the dataset should persist in the database after the session terminates
overwrite (False) – whether to overwrite any existing files if the dataset is to be downloaded
cleanup (True) – whether to cleanup any temporary files generated during download
progress (None) – whether to render a progress bar (True/False), use the default value fiftyone.config.show_progress_bars (None), or a progress callback function to invoke instead
**kwargs – optional arguments to pass to the fiftyone.utils.data.importers.DatasetImporter constructor. If download_if_necessary == True, then kwargs can also contain arguments for download_zoo_dataset()

Returns

a fiftyone.core.dataset.Dataset

fiftyone.zoo.datasets.find_zoo_dataset(name, split=None)¶

Returns the directory containing the given zoo dataset.

If a split is provided, the path to the dataset split is returned; otherwise, the path to the root directory is returned.

The dataset must be downloaded. Use download_zoo_dataset() to download datasets.

Parameters

name – the name of the zoo dataset
split (None) –

Returns

the directory containing the dataset

Raises

ValueError – if the dataset or split does not exist or has not been downloaded

fiftyone.zoo.datasets.load_zoo_dataset_info(name, dataset_dir=None)¶

Loads the ZooDatasetInfo for the specified zoo dataset.

The dataset must be downloaded. Use download_zoo_dataset() to download datasets.

Parameters

name – the name of the zoo dataset
dataset_dir (None) – the directory in which the dataset is stored. By default, the dataset is located in fiftyone.config.dataset_zoo_dir

Returns

the ZooDatasetInfo for the dataset

Raises

ValueError – if the dataset has not been downloaded

fiftyone.zoo.datasets.get_zoo_dataset(name, **kwargs)¶

Returns the ZooDataset instance for the dataset with the given name.

If the dataset is available from multiple sources, the default source is used.

Parameters

name – the name of the zoo dataset
**kwargs – optional arguments for ZooDataset

Returns

the ZooDataset instance

fiftyone.zoo.datasets.delete_zoo_dataset(name, split=None)¶

Deletes the zoo dataset from local disk, if necessary.

If a split is provided, only that split is deleted.

Parameters

name – the name of the zoo dataset
split (None) –

class fiftyone.zoo.datasets.ZooDatasetInfo(zoo_dataset, dataset_type, num_samples, downloaded_splits=None, parameters=None, classes=None)¶

Bases: eta.core.serial.Serializable

Class containing info about a dataset in the FiftyOne Dataset Zoo.

Parameters

zoo_dataset – the ZooDataset instance for the dataset
dataset_type – the fiftyone.types.Dataset type of the dataset
num_samples – the total number of samples in all downloaded splits of the dataset
downloaded_splits (None) – a dict of ZooDatasetSplitInfo instances describing the downloaded splits of the dataset, if applicable
parameters (None) – a dict of parameters for the dataset
classes (None) – a list of class label strings

Attributes:

`name`	The name of the dataset.
`zoo_dataset`	The fully-qualified class string for the `ZooDataset` of the dataset.
`dataset_type`	The fully-qualified class string of the `fiftyone.types.Dataset` type.
`supported_splits`	A tuple of supported splits for the dataset, or None if the dataset does not have splits.

Methods:

`get_zoo_dataset`()	Returns the `ZooDataset` instance for the dataset.
`get_dataset_type`()	Returns the `fiftyone.types.Dataset` type instance for the dataset.
`is_split_downloaded`(split)	Whether the given dataset split is downloaded.
`add_split`(split_info)	Adds the split to the dataset.
`remove_split`(split)	Removes the split from the dataset.
`attributes`()	Returns a list of class attributes to be serialized.
`from_dict`(d)	Loads a `ZooDatasetInfo` from a JSON dictionary.
`from_json`(json_path[, upgrade, warn_deprecated])	Loads a `ZooDatasetInfo` from a JSON file on disk.
`copy`()	Returns a deep copy of the object.
`custom_attributes`([dynamic, private])	Returns a customizable list of class attributes.
`from_str`(s, args, *kwargs)	Constructs a Serializable object from a JSON string.
`get_class_name`()	Returns the fully-qualified class name string of this object.
`serialize`([reflective])	Serializes the object into a dictionary.
`to_str`([pretty_print])	Returns a string representation of this object.
`write_json`(path[, pretty_print])	Serializes the object and writes it to disk.

property name¶: The name of the dataset.

property zoo_dataset¶: The fully-qualified class string for the ZooDataset of the dataset.

property dataset_type¶: The fully-qualified class string of the fiftyone.types.Dataset type.

property supported_splits¶: A tuple of supported splits for the dataset, or None if the dataset does not have splits.

get_zoo_dataset()¶

Returns the ZooDataset instance for the dataset.

Returns: a ZooDataset instance

get_dataset_type()¶

Returns the fiftyone.types.Dataset type instance for the dataset.

Returns: a fiftyone.types.Dataset instance

is_split_downloaded(split)¶

Whether the given dataset split is downloaded.

Parameters: split – the dataset split
Returns: True/False

add_split(split_info)¶

Adds the split to the dataset.

Parameters: split_info – a ZooDatasetSplitInfo

remove_split(split)¶

Removes the split from the dataset.

Parameters: split – the name of the split

attributes()¶

Returns a list of class attributes to be serialized.

Returns: a list of class attributes

classmethod from_dict(d)¶

Loads a ZooDatasetInfo from a JSON dictionary.

Parameters: d – a JSON dictionary
Returns: a ZooDatasetInfo

classmethod from_json(json_path, upgrade=False, warn_deprecated=False)¶

Loads a ZooDatasetInfo from a JSON file on disk.

Parameters

json_path – path to JSON file
upgrade (False) – whether to upgrade the JSON file on disk if any migrations were necessary
warn_deprecated (False) – whether to issue a warning if the dataset has a deprecated format

Returns

a ZooDatasetInfo

copy()¶

Returns a deep copy of the object.

Returns: a Serializable instance

custom_attributes(dynamic=False, private=False)¶

Returns a customizable list of class attributes.

By default, all attributes in vars(self) are returned, minus private attributes (those starting with “_”).

Parameters

dynamic – whether to include dynamic properties, e.g., those defined by getter/setter methods or the @property decorator. By default, this is False
private – whether to include private properties, i.e., those starting with “_”. By default, this is False

Returns

a list of class attributes

classmethod from_str(s, *args, **kwargs)¶

Constructs a Serializable object from a JSON string.

Subclasses may override this method, but, by default, this method simply parses the string and calls from_dict(), which subclasses must implement.

Parameters

s – a JSON string representation of a Serializable object
*args – optional positional arguments for self.from_dict()
**kwargs – optional keyword arguments for self.from_dict()

Returns

an instance of the Serializable class

classmethod get_class_name()¶: Returns the fully-qualified class name string of this object.

serialize(reflective=False)¶

Serializes the object into a dictionary.

Serialization is applied recursively to all attributes in the object, including element-wise serialization of lists and dictionary values.

Parameters: reflective – whether to include reflective attributes when serializing the object. By default, this is False
Returns: a JSON dictionary representation of the object

to_str(pretty_print=True, **kwargs)¶

Returns a string representation of this object.

Parameters

pretty_print – whether to render the JSON in human readable format with newlines and indentations. By default, this is True
**kwargs – optional keyword arguments for self.serialize()

Returns

a string representation of the object

write_json(path, pretty_print=False, **kwargs)¶

Serializes the object and writes it to disk.

Parameters

path – the output path
pretty_print – whether to render the JSON in human readable format with newlines and indentations. By default, this is False
**kwargs – optional keyword arguments for self.serialize()

class fiftyone.zoo.datasets.ZooDatasetSplitInfo(split, num_samples)¶

Bases: eta.core.serial.Serializable

Class containing info about a split of a dataset in the FiftyOne Dataset Zoo.

Parameters

split – the name of the split
num_samples – the number of samples in the split

Methods:

`attributes`()	Returns a list of class attributes to be serialized.
`from_dict`(d)	Loads a `ZooDatasetSplitInfo` from a JSON dictionary.
`copy`()	Returns a deep copy of the object.
`custom_attributes`([dynamic, private])	Returns a customizable list of class attributes.
`from_json`(path, args, *kwargs)	Constructs a Serializable object from a JSON file.
`from_str`(s, args, *kwargs)	Constructs a Serializable object from a JSON string.
`get_class_name`()	Returns the fully-qualified class name string of this object.
`serialize`([reflective])	Serializes the object into a dictionary.
`to_str`([pretty_print])	Returns a string representation of this object.
`write_json`(path[, pretty_print])	Serializes the object and writes it to disk.

attributes()¶

Returns a list of class attributes to be serialized.

Returns: a list of class attributes

classmethod from_dict(d)¶

Loads a ZooDatasetSplitInfo from a JSON dictionary.

Parameters: d – a JSON dictionary
Returns: a ZooDatasetSplitInfo

copy()¶

Returns a deep copy of the object.

Returns: a Serializable instance

custom_attributes(dynamic=False, private=False)¶

Returns a customizable list of class attributes.

By default, all attributes in vars(self) are returned, minus private attributes (those starting with “_”).

Parameters

dynamic – whether to include dynamic properties, e.g., those defined by getter/setter methods or the @property decorator. By default, this is False
private – whether to include private properties, i.e., those starting with “_”. By default, this is False

Returns

a list of class attributes

classmethod from_json(path, *args, **kwargs)¶

Constructs a Serializable object from a JSON file.

Subclasses may override this method, but, by default, this method simply reads the JSON and calls from_dict(), which subclasses must implement.

Parameters

path – the path to the JSON file on disk
*args – optional positional arguments for self.from_dict()
**kwargs – optional keyword arguments for self.from_dict()

Returns

an instance of the Serializable class

classmethod from_str(s, *args, **kwargs)¶

Constructs a Serializable object from a JSON string.

Subclasses may override this method, but, by default, this method simply parses the string and calls from_dict(), which subclasses must implement.

Parameters

s – a JSON string representation of a Serializable object
*args – optional positional arguments for self.from_dict()
**kwargs – optional keyword arguments for self.from_dict()

Returns

an instance of the Serializable class

classmethod get_class_name()¶: Returns the fully-qualified class name string of this object.

serialize(reflective=False)¶

Serializes the object into a dictionary.

Serialization is applied recursively to all attributes in the object, including element-wise serialization of lists and dictionary values.

Parameters: reflective – whether to include reflective attributes when serializing the object. By default, this is False
Returns: a JSON dictionary representation of the object

to_str(pretty_print=True, **kwargs)¶

Returns a string representation of this object.

Parameters

pretty_print – whether to render the JSON in human readable format with newlines and indentations. By default, this is True
**kwargs – optional keyword arguments for self.serialize()

Returns

a string representation of the object

write_json(path, pretty_print=False, **kwargs)¶

Serializes the object and writes it to disk.

Parameters

path – the output path
pretty_print – whether to render the JSON in human readable format with newlines and indentations. By default, this is False
**kwargs – optional keyword arguments for self.serialize()

class fiftyone.zoo.datasets.ZooDataset¶

Bases: object

Base class for datasets made available in the FiftyOne Dataset Zoo.

Attributes:

`name`	The name of the dataset.
`tags`	A tuple of tags for the dataset.
`has_tags`	Whether the dataset has tags.
`parameters`	An optional dict of parameters describing the configuration of the zoo dataset when it was downloaded.
`supported_splits`	A tuple of supported splits for the dataset, or None if the dataset does not have splits.
`has_splits`	Whether the dataset has splits.
`has_patches`	Whether the dataset has patches that may need to be applied to already downloaded files.
`supports_partial_downloads`	Whether the dataset supports downloading partial subsets of its splits.
`requires_manual_download`	Whether this dataset requires some files to be manually downloaded by the user before the dataset can be loaded.
`importer_kwargs`	A dict of default kwargs to pass to this dataset’s `fiftyone.utils.data.importers.DatasetImporter`.

Methods:

`has_tag`(tag)	Whether the dataset has the given tag.
`has_split`(split)	Whether the dataset has the given split.
`get_split_dir`(dataset_dir, split)	Returns the directory for the given split of the dataset.
`load_info`(dataset_dir[, upgrade, …])	Loads the `ZooDatasetInfo` from the given dataset directory.
`get_info_path`(dataset_dir)	Returns the path to the `ZooDatasetInfo` for the dataset.
`download_and_prepare`([dataset_dir, split, …])	Downloads the dataset and prepares it for use.

property name¶: The name of the dataset.

property tags¶: A tuple of tags for the dataset.

property has_tags¶: Whether the dataset has tags.

property parameters¶: An optional dict of parameters describing the configuration of the zoo dataset when it was downloaded.

property supported_splits¶: A tuple of supported splits for the dataset, or None if the dataset does not have splits.

property has_splits¶: Whether the dataset has splits.

property has_patches¶: Whether the dataset has patches that may need to be applied to already downloaded files.

property supports_partial_downloads¶: Whether the dataset supports downloading partial subsets of its splits.

property requires_manual_download¶: Whether this dataset requires some files to be manually downloaded by the user before the dataset can be loaded.

property importer_kwargs¶: A dict of default kwargs to pass to this dataset’s fiftyone.utils.data.importers.DatasetImporter.

has_tag(tag)¶

Whether the dataset has the given tag.

Parameters: tag – the tag
Returns: True/False

has_split(split)¶

Whether the dataset has the given split.

Parameters: split – the dataset split
Returns: True/False

get_split_dir(dataset_dir, split)¶

Returns the directory for the given split of the dataset.

Parameters

dataset_dir – the dataset directory
split – the dataset split

Returns

the directory that will/does hold the specified split

static load_info(dataset_dir, upgrade=True, warn_deprecated=False)¶

Loads the ZooDatasetInfo from the given dataset directory.

Parameters

dataset_dir – the directory in which to construct the dataset
upgrade (True) – whether to upgrade the JSON file on disk if any migrations were necessary
warn_deprecated (False) – whether to issue a warning if the dataset has a deprecated format

Returns

the ZooDatasetInfo for the dataset

static get_info_path(dataset_dir)¶

Returns the path to the ZooDatasetInfo for the dataset.

Parameters: dataset_dir – the dataset directory
Returns: the path to the ZooDatasetInfo

download_and_prepare(dataset_dir=None, split=None, splits=None, overwrite=False, cleanup=True)¶

Downloads the dataset and prepares it for use.

If the requested splits have already been downloaded, they are not re-downloaded unless overwrite is True.

Parameters

dataset_dir (None) – the directory in which to construct the dataset. By default, it is written to a subdirectory of fiftyone.config.dataset_zoo_dir
split (None) – split nor splits are provided, the full dataset is downloaded
splits (None) – a list of splits to download, if applicable. If neither split nor splits are provided, the full dataset is downloaded
overwrite (False) – whether to overwrite any existing files
cleanup (True) – whether to cleanup any temporary files generated during download

Returns

tuple of

info: the ZooDatasetInfo for the dataset
dataset_dir: the directory containing the dataset

class fiftyone.zoo.datasets.DeprecatedZooDataset¶

Bases: fiftyone.zoo.datasets.ZooDataset

Class representing a zoo dataset that no longer exists in the FiftyOne Dataset Zoo.

Attributes:

`name`	The name of the dataset.
`supported_splits`	A tuple of supported splits for the dataset, or None if the dataset does not have splits.
`has_patches`	Whether the dataset has patches that may need to be applied to already downloaded files.
`has_splits`	Whether the dataset has splits.
`has_tags`	Whether the dataset has tags.
`importer_kwargs`	A dict of default kwargs to pass to this dataset’s `fiftyone.utils.data.importers.DatasetImporter`.
`parameters`	An optional dict of parameters describing the configuration of the zoo dataset when it was downloaded.
`requires_manual_download`	Whether this dataset requires some files to be manually downloaded by the user before the dataset can be loaded.
`supports_partial_downloads`	Whether the dataset supports downloading partial subsets of its splits.
`tags`	A tuple of tags for the dataset.

Methods:

`download_and_prepare`([dataset_dir, split, …])	Downloads the dataset and prepares it for use.
`get_info_path`(dataset_dir)	Returns the path to the `ZooDatasetInfo` for the dataset.
`get_split_dir`(dataset_dir, split)	Returns the directory for the given split of the dataset.
`has_split`(split)	Whether the dataset has the given split.
`has_tag`(tag)	Whether the dataset has the given tag.
`load_info`(dataset_dir[, upgrade, …])	Loads the `ZooDatasetInfo` from the given dataset directory.

property name¶: The name of the dataset.

property supported_splits¶: A tuple of supported splits for the dataset, or None if the dataset does not have splits.

download_and_prepare(dataset_dir=None, split=None, splits=None, overwrite=False, cleanup=True)¶

Downloads the dataset and prepares it for use.

If the requested splits have already been downloaded, they are not re-downloaded unless overwrite is True.

Parameters

dataset_dir (None) – the directory in which to construct the dataset. By default, it is written to a subdirectory of fiftyone.config.dataset_zoo_dir
split (None) – split nor splits are provided, the full dataset is downloaded
splits (None) – a list of splits to download, if applicable. If neither split nor splits are provided, the full dataset is downloaded
overwrite (False) – whether to overwrite any existing files
cleanup (True) – whether to cleanup any temporary files generated during download

Returns

tuple of

info: the ZooDatasetInfo for the dataset
dataset_dir: the directory containing the dataset

static get_info_path(dataset_dir)¶

Returns the path to the ZooDatasetInfo for the dataset.

Parameters: dataset_dir – the dataset directory
Returns: the path to the ZooDatasetInfo

get_split_dir(dataset_dir, split)¶

Returns the directory for the given split of the dataset.

Parameters

dataset_dir – the dataset directory
split – the dataset split

Returns

the directory that will/does hold the specified split

property has_patches¶: Whether the dataset has patches that may need to be applied to already downloaded files.

has_split(split)¶

Whether the dataset has the given split.

Parameters: split – the dataset split
Returns: True/False

property has_splits¶: Whether the dataset has splits.

has_tag(tag)¶

Whether the dataset has the given tag.

Parameters: tag – the tag
Returns: True/False

property has_tags¶: Whether the dataset has tags.

property importer_kwargs¶: A dict of default kwargs to pass to this dataset’s fiftyone.utils.data.importers.DatasetImporter.

static load_info(dataset_dir, upgrade=True, warn_deprecated=False)¶

Loads the ZooDatasetInfo from the given dataset directory.

Parameters

dataset_dir – the directory in which to construct the dataset
upgrade (True) – whether to upgrade the JSON file on disk if any migrations were necessary
warn_deprecated (False) – whether to issue a warning if the dataset has a deprecated format

Returns

the ZooDatasetInfo for the dataset

property parameters¶: An optional dict of parameters describing the configuration of the zoo dataset when it was downloaded.

property requires_manual_download¶: Whether this dataset requires some files to be manually downloaded by the user before the dataset can be loaded.

property supports_partial_downloads¶: Whether the dataset supports downloading partial subsets of its splits.

property tags¶: A tuple of tags for the dataset.