fiftyone.utils.kinetics

module documentation

(source)

Utilities for working with the Kinetics dataset.

Class	`ClasswiseS3KineticsDatasetInfo`	Undocumented
Class	`Kinetics400DatasetInfo`	Kinetics 400-specific dataset info.
Class	`Kinetics600DatasetInfo`	Kinetics 600-specific dataset info.
Class	`Kinetics7002020DatasetInfo`	Kinetics 700-2020-specific dataset info.
Class	`Kinetics700DatasetInfo`	Kinetics 700-specific dataset info.
Class	`KineticsDatasetDownloader`	Clas that downloads and extracts Kinetics tars from AWS.
Class	`KineticsDatasetInfo`	Class that contains information such as paths, labels, and sample IDs for a Kinetics download.
Class	`KineticsDatasetManager`	Class that manages the sample IDs and labels that need to be downloaded as well as performing the actual downloading.
Class	`KineticsDownloadConfig`	Config class for a Kinetics download run.
Function	`download_kinetics_split`	Utility that downloads full or partial splits of the Kinetics dataset.
Variable	`logger`	Undocumented
Function	`_flatten_list`	Undocumented
Constant	`_ANNOTATION_DOWNLOAD_LINKS`	Undocumented
Constant	`_INFO_VERSION_MAP`	Undocumented
Constant	`_SPLIT_MAP`	Undocumented

def download_kinetics_split(dataset_dir, split, classes=None, num_workers=None, shuffle=None, seed=None, max_samples=None, retry_errors=False, scratch_dir=None, version='700-2020'): (source) ¶

Utility that downloads full or partial splits of the Kinetics dataset.

The downloaded splits are stored on disk in :ref:`VideoClassificationDirectoryTree format <VideoClassificationDirectoryTree-import>`.

Parameters
dataset_dir	the directory to download the dataset
split	the split to download. Supported values are `("train", "validation", "test")`
classes:`None`	a string or list of strings specifying required classes to load. If provided, only samples containing at least one instance of a specified class will be loaded
num_workers:`None`	a suggested number of threads to use when downloading individual videos
shuffle:`False`	whether to randomly shuffle the order in which samples are chosen for partial downloads
seed:`None`	a random seed to use when shuffling
max_samples:`None`	a maximum number of samples to load per split. If `classes` are also specified, only up to the number of samples that contain at least one specified class will be loaded. By default, all matching samples are loaded
retry_errors:`False`	whether to retry downloading samples from YouTube that have previously raised an error
scratch_dir:`None`	a scratch directory to use to store temporary files
version:"700-2020"	the version of the Kinetics dataset to download ("400", "600", "700", or "700-2020")
Returns
a tuple of	num_samples: the total number of downloaded videos, or `None` if everything was already downloaded classes: the list of all classes, or `None` if everything was already downloaded did_download: whether any content was downloaded (True) or if all necessary files were already downloaded (False)

logger = (source) ¶

Undocumented

def _flatten_list(l): (source) ¶

Undocumented

_ANNOTATION_DOWNLOAD_LINKS: dict[str, str] = (source) ¶

Undocumented

Value

{'400': 'https://storage.googleapis.com/deepmind-media/Datasets/kinetics400.tar.↵
gz',
 '600': 'https://storage.googleapis.com/deepmind-media/Datasets/kinetics600.tar.↵
gz',
 '700': 'https://storage.googleapis.com/deepmind-media/Datasets/kinetics700.tar.↵
gz',
 '700-2020': 'https://storage.googleapis.com/deepmind-media/Datasets/kinetics700↵
...

_INFO_VERSION_MAP = (source) ¶

Undocumented

Value

{'400': Kinetics400DatasetInfo,
 '600': Kinetics600DatasetInfo,
 '700': Kinetics700DatasetInfo,
 '700-2020': Kinetics7002020DatasetInfo}

_SPLIT_MAP: dict[str, str] = (source) ¶

Undocumented

Value

{'test': 'test', 'train': 'train', 'validation': 'validate'}