FiftyOne Command-Line Interface (CLI)¶
Installing FiftyOne automatically installs fiftyone, a command-line interface
(CLI) for interacting with FiftyOne. This utility provides access to many
useful features, including creating and inspecting datasets, visualizing
datasets in the App, exporting datasets and converting dataset formats,
and downloading datasets from the FiftyOne Dataset Zoo.
Quickstart¶
To see the available top-level commands, type:
fiftyone --help
You can learn more about any available subcommand via:
fiftyone <command> --help
For example, to see your current FiftyOne config, you can execute
fiftyone config.
Tab completion¶
To enable tab completion in bash, add the following line to your ~/.bashrc:
eval "$(register-python-argcomplete fiftyone)"
To enable tab completion in zsh, add these lines to your ~/.zshrc:
autoload bashcompinit
bashcompinit
eval "$(register-python-argcomplete fiftyone)"
To enable tab completion in tcsh, add these lines to your ~/.tcshrc:
eval `register-python-argcomplete --shell tcsh fiftyone`
FiftyOne CLI¶
The FiftyOne command-line interface.
fiftyone [-h] [-v] [--all-help]
{quickstart,annotation,brain,evaluation,app,config,constants,convert,datasets,migrate,operators,delegated,plugins,utils,zoo}
...
Arguments
optional arguments:
-h, --help show this help message and exit
-v, --version show version info
--all-help show help recursively and exit
available commands:
{quickstart,annotation,brain,evaluation,app,config,constants,convert,datasets,migrate,operators,delegated,plugins,utils,zoo}
quickstart Launch a FiftyOne quickstart.
annotation Tools for working with the FiftyOne annotation API.
brain Tools for working with the FiftyOne Brain.
evaluation Tools for working with the FiftyOne evaluation API.
app Tools for working with the FiftyOne App.
config Tools for working with your FiftyOne config.
constants Print constants from `fiftyone.constants`.
convert Convert datasets on disk between supported formats.
datasets Tools for working with FiftyOne datasets.
migrate Tools for migrating the FiftyOne database.
operators Tools for working with FiftyOne operators.
delegated Tools for working with FiftyOne delegated operations.
plugins Tools for working with FiftyOne plugins.
utils FiftyOne utilities.
zoo Tools for working with the FiftyOne Zoo.
FiftyOne quickstart¶
Launch a FiftyOne quickstart.
fiftyone quickstart [-h] [-v] [-p PORT] [-A ADDRESS] [-r] [-a] [-w WAIT]
Arguments
optional arguments:
-h, --help show this help message and exit
-v, --video launch the quickstart with a video dataset
-p PORT, --port PORT the port number to use
-A ADDRESS, --address ADDRESS
the address (server name) to use
-r, --remote whether to launch a remote App session
-w WAIT, --wait WAIT the number of seconds to wait for a new App
connection before returning if all connections are
lost. If negative, the process will wait forever,
regardless of connections
Examples
# Launch the quickstart
fiftyone quickstart
# Launch the quickstart with a video dataset
fiftyone quickstart --video
# Launch the quickstart as a remote session
fiftyone quickstart --remote
FiftyOne config¶
Tools for working with your FiftyOne config.
fiftyone config [-h] [-l] [FIELD]
Arguments
positional arguments:
FIELD a config field to print
optional arguments:
-h, --help show this help message and exit
-l, --locate print the location of your config on disk
Examples
# Print your entire config
fiftyone config
# Print a specific config field
fiftyone config <field>
# Print the location of your config on disk (if one exists)
fiftyone config --locate
Print constants¶
Print constants from fiftyone.constants.
fiftyone constants [-h] [CONSTANT]
Arguments
positional arguments:
CONSTANT the constant to print
optional arguments:
-h, --help show this help message and exit
Examples
# Print all constants
fiftyone constants
# Print a specific constant
fiftyone constants <CONSTANT>
Convert dataset formats¶
Convert datasets on disk between supported formats.
fiftyone convert [-h] --input-type INPUT_TYPE --output-type OUTPUT_TYPE
[--input-dir INPUT_DIR]
[--input-kwargs KEY=VAL [KEY=VAL ...]]
[--output-dir OUTPUT_DIR]
[--output-kwargs KEY=VAL [KEY=VAL ...]] [-o]
Arguments
optional arguments:
-h, --help show this help message and exit
--input-dir INPUT_DIR
the directory containing the dataset
--input-kwargs KEY=VAL [KEY=VAL ...]
additional keyword arguments for
`fiftyone.utils.data.convert_dataset(..., input_kwargs=)`
--output-dir OUTPUT_DIR
the directory to which to write the output dataset
--output-kwargs KEY=VAL [KEY=VAL ...]
additional keyword arguments for
`fiftyone.utils.data.convert_dataset(..., output_kwargs=)`
-o, --overwrite whether to overwrite an existing output directory
required arguments:
--input-type INPUT_TYPE
the fiftyone.types.Dataset type of the input dataset
--output-type OUTPUT_TYPE
the fiftyone.types.Dataset type to output
Examples
# Convert an image classification directory tree to TFRecords format
fiftyone convert \
--input-dir /path/to/image-classification-directory-tree \
--input-type fiftyone.types.ImageClassificationDirectoryTree \
--output-dir /path/for/tf-image-classification-dataset \
--output-type fiftyone.types.TFImageClassificationDataset
# Convert a COCO detection dataset to CVAT image format
fiftyone convert \
--input-dir /path/to/coco-detection-dataset \
--input-type fiftyone.types.COCODetectionDataset \
--output-dir /path/for/cvat-image-dataset \
--output-type fiftyone.types.CVATImageDataset
# Perform a customized conversion via optional kwargs
fiftyone convert \
--input-dir /path/to/coco-detection-dataset \
--input-type fiftyone.types.COCODetectionDataset \
--input-kwargs max_samples=100 shuffle=True \
--output-dir /path/for/cvat-image-dataset \
--output-type fiftyone.types.TFObjectDetectionDataset \
--output-kwargs force_rgb=True \
--overwrite
FiftyOne datasets¶
Tools for working with FiftyOne datasets.
fiftyone datasets [-h] [--all-help]
{list,info,create,head,tail,stream,export,delete} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{list,info,create,head,tail,stream,export,delete}
list List FiftyOne datasets.
info Print information about FiftyOne datasets.
stats Print stats about FiftyOne datasets on disk.
create Tools for creating FiftyOne datasets.
head Prints the first few samples in a FiftyOne dataset.
tail Prints the last few samples in a FiftyOne dataset.
stream Streams the samples in a FiftyOne dataset.
export Export FiftyOne datasets to disk in supported formats.
draw Writes annotated versions of samples in FiftyOne datasets to disk.
rename Rename FiftyOne datasets.
delete Delete FiftyOne datasets.
List datasets¶
List FiftyOne datasets.
fiftyone datasets list [-h] [-p PATT] [-t TAG [TAG ...]]
Arguments
optional arguments:
-h, --help show this help message and exit
-p PATT, --glob-patt PATT
an optional glob pattern of dataset names to include
-t TAG [TAG ...], --tags TAG [TAG ...]
only show datasets with the given tag(s)
Examples
# List available datasets
fiftyone datasets list
# List datasets matching a given pattern
fiftyone datasets list --glob-patt 'quickstart-*'
# List datasets with the given tag(s)
fiftyone datasets list --tags automotive healthcare
Print dataset information¶
Print information about FiftyOne datasets.
fiftyone datasets info [-h] [-p PATT] [-t TAG [TAG ...]] [-s FIELD] [-r] [NAME]
Arguments
positional arguments:
NAME the name of a dataset
optional arguments:
-h, --help show this help message and exit
-p PATT, --glob-patt PATT
an optional glob pattern of dataset names to include
-t TAG [TAG ...], --tags TAG [TAG ...]
only show datasets with the given tag(s)
-s FIELD, --sort-by FIELD
a field to sort the dataset rows by
-r, --reverse whether to print the results in reverse order
Examples
# Print basic information about multiple datasets
fiftyone datasets info
fiftyone datasets info --glob-patt 'quickstart-*'
fiftyone datasets info --tags automotive healthcare
fiftyone datasets info --sort-by created_at
fiftyone datasets info --sort-by name --reverse
# Print information about a specific dataset
fiftyone datasets info <name>
Print dataset stats¶
Print stats about FiftyOne datasets on disk.
fiftyone datasets stats [-h] [-m] [-c] NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
-m, --include-media whether to include stats about the size of the raw
media in the dataset
-c, --compressed whether to return the sizes of collections in their
compressed form on disk
Examples
# Print stats about the given dataset on disk
fiftyone datasets stats <name>
Create datasets¶
Tools for creating FiftyOne datasets.
fiftyone datasets create [-h] [-n NAME] [-d DATASET_DIR] [-j JSON_PATH]
[-t TYPE] [-k KEY=VAL [KEY=VAL ...]]
Arguments
optional arguments:
-h, --help show this help message and exit
-n NAME, --name NAME a name for the dataset
-d DATASET_DIR, --dataset-dir DATASET_DIR
the directory containing the dataset
-j JSON_PATH, --json-path JSON_PATH
the path to a samples JSON file to load
-t TYPE, --type TYPE the fiftyone.types.Dataset type of the dataset
-k KEY=VAL [KEY=VAL ...], --kwargs KEY=VAL [KEY=VAL ...]
additional type-specific keyword arguments for
`fiftyone.core.dataset.Dataset.from_dir()`
Examples
# Create a dataset from the given data on disk
fiftyone datasets create \
--name <name> --dataset-dir <dataset-dir> --type <type>
# Create a dataset from a random subset of the data on disk
fiftyone datasets create \
--name <name> --dataset-dir <dataset-dir> --type <type> \
--kwargs max_samples=50 shuffle=True
# Create a dataset from the given samples JSON file
fiftyone datasets create --json-path <json-path>
Print dataset head¶
Prints the first few samples in a FiftyOne dataset.
fiftyone datasets head [-h] [-n NUM_SAMPLES] NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
-n NUM_SAMPLES, --num-samples NUM_SAMPLES
the number of samples to print
Examples
# Prints the first few samples in a dataset
fiftyone datasets head <name>
# Prints the given number of samples from the head of a dataset
fiftyone datasets head <name> --num-samples <num-samples>
Print dataset tail¶
Prints the last few samples in a FiftyOne dataset.
fiftyone datasets tail [-h] [-n NUM_SAMPLES] NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
-n NUM_SAMPLES, --num-samples NUM_SAMPLES
the number of samples to print
Examples
# Print the last few samples in a dataset
fiftyone datasets tail <name>
# Print the given number of samples from the tail of a dataset
fiftyone datasets tail <name> --num-samples <num-samples>
Stream samples to the terminal¶
Stream samples in a FiftyOne dataset to the terminal.
fiftyone datasets stream [-h] NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
Examples
# Stream the samples of the dataset to the terminal
fiftyone datasets stream <name>
Export datasets¶
Export FiftyOne datasets to disk in supported formats.
fiftyone datasets export [-h] [-d EXPORT_DIR] [-j JSON_PATH]
[-f LABEL_FIELD] [-t TYPE]
[--filters KEY=VAL [KEY=VAL ...]]
[-k KEY=VAL [KEY=VAL ...]]
NAME
Arguments
positional arguments:
NAME the name of the dataset to export
optional arguments:
-h, --help show this help message and exit
-d EXPORT_DIR, --export-dir EXPORT_DIR
the directory in which to export the dataset
-j JSON_PATH, --json-path JSON_PATH
the path to export the dataset in JSON format
-f LABEL_FIELD, --label-field LABEL_FIELD
the name of the label field to export
-t TYPE, --type TYPE the fiftyone.types.Dataset type in which to export
--filters KEY=VAL [KEY=VAL ...]
specific sample tags or class labels to export. To
use sample tags, pass tags as `tags=train,val` and
to use label filters, pass label field and values
as in ground_truth=car,person,dog
-k KEY=VAL [KEY=VAL ...], --kwargs KEY=VAL [KEY=VAL ...]
additional type-specific keyword arguments for
`fiftyone.core.collections.SampleCollection.export()`
Examples
# Export the dataset to disk in the specified format
fiftyone datasets export <name> \
--export-dir <export-dir> --type <type> --label-field <label-field>
# Export the dataset to disk in JSON format
fiftyone datasets export <name> --json-path <json-path>
# Only export cats and dogs from the validation split
fiftyone datasets export <name> \\
--filters tags=validation ground_truth=cat,dog \\
--export-dir <export-dir> --type <type> --label-field ground_truth
# Perform a customized export of a dataset
fiftyone datasets export <name> \
--type <type> \
--kwargs labels_path=/path/for/labels.json
Drawing labels on samples¶
Renders annotated versions of samples in FiftyOne datasets to disk.
fiftyone datasets draw [-h] [-d OUTPUT_DIR] [-f LABEL_FIELDS] NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
-d OUTPUT_DIR, --output-dir OUTPUT_DIR
the directory to write the annotated media
-f LABEL_FIELDS, --label-fields LABEL_FIELDS
a comma-separated list of label fields to export
Examples
# Write annotated versions of the media in the dataset with the
# specified label field(s) overlaid to disk
fiftyone datasets draw <name> \
--output-dir <output-dir> --label-fields <list>,<of>,<fields>
Rename datasets¶
Rename FiftyOne datasets.
fiftyone datasets rename [-h] NAME NEW_NAME
Arguments
positional arguments:
NAME the name of the dataset
NEW_NAME a new name for the dataset
optional arguments:
-h, --help show this help message and exit
Examples
# Rename the dataset
fiftyone datasets rename <old-name> <new-name>
Delete datasets¶
Delete FiftyOne datasets.
fiftyone datasets delete [-h] [-g GLOB_PATT] [--non-persistent]
[NAME [NAME ...]]
Arguments
positional arguments:
NAME the dataset name(s) to delete
optional arguments:
-h, --help show this help message and exit
-g GLOB_PATT, --glob-patt GLOB_PATT
a glob pattern of datasets to delete
--non-persistent delete all non-persistent datasets
Examples
# Delete the datasets with the given name(s)
fiftyone datasets delete <name1> <name2> ...
# Delete the datasets whose names match the given glob pattern
fiftyone datasets delete --glob-patt <glob-patt>
# Delete all non-persistent datasets
fiftyone datasets delete --non-persistent
FiftyOne migrations¶
Tools for migrating the FiftyOne database.
See this page for more information about migrating FiftyOne deployments.
fiftyone migrate [-h] [-i] [-a]
[-v VERSION]
[-n DATASET_NAME [DATASET_NAME ...]]
[--error-level LEVEL]
[--verbose]
Arguments
optional arguments:
-h, --help show this help message and exit
-i, --info whether to print info about the current revisions
-a, --all whether to migrate the database and all datasets
-v VERSION, --version VERSION
the revision to migrate to
-n DATASET_NAME [DATASET_NAME ...], --dataset-name DATASET_NAME [DATASET_NAME ...]
the name of a specific dataset to migrate
--error-level LEVEL the error level (0=error, 1=warn, 2=ignore) to use
when migrating individual datasets
--verbose whether to log incremental migrations that are performed
Examples
# Print information about the current revisions of all datasets
fiftyone migrate --info
# Migrate the database and all datasets to the current client version
fiftyone migrate --all
# Migrate to a specific revision
fiftyone migrate --all --version <VERSION>
# Migrate a specific dataset
fiftyone migrate ... --dataset-name <DATASET_NAME>
# Update the database version without migrating any existing datasets
fiftyone migrate
FiftyOne operators¶
Tools for working with FiftyOne operators and panels.
fiftyone operators [-h] [--all-help] {list,info} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{list,info}
list List operators and panels that you've installed locally.
info Prints information about operators and panels that you've installed locally.
List operators¶
List operators and panels that you’ve installed locally.
fiftyone operators list [-h] [-e] [-d] [-o] [-p] [-n]
Arguments
optional arguments:
-h, --help show this help message and exit
-e, --enabled only show enabled operators and panels
-d, --disabled only show disabled operators and panels
-o, --operators-only only show operators
-p, --panels-only only show panels
-n, --names-only only show names
Examples
# List all available operators and panels
fiftyone operators list
# List enabled operators and panels
fiftyone operators list --enabled
# List disabled operators and panels
fiftyone operators list --disabled
# Only list panels
fiftyone operators list --panels-only
Operator info¶
Prints information about operators and panels that you’ve installed locally.
fiftyone operators info [-h] URI
Arguments
positional arguments:
URI the operator or panel URI
optional arguments:
-h, --help show this help message and exit
Examples
# Prints information about an operator or panel
fiftyone operators info <uri>
FiftyOne delegated operations¶
Tools for working with FiftyOne delegated operations.
fiftyone delegated [-h] [--all-help] {launch,list,info,fail,delete,cleanup} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{launch,list,info,fail,delete,cleanup}
launch Launches a service for running delegated operations.
list List delegated operations.
info Prints information about a delegated operation.
fail Manually mark delegated as failed.
delete Delete delegated operations.
cleanup Cleanup delegated operations.
Launch delegated service¶
Launches a service for running delegated operations.
fiftyone delegated launch [-h] [-t TYPE]
Arguments
optional arguments:
-h, --help show this help message and exit
-t TYPE, --type TYPE the type of service to launch. The default is 'local'
Examples
# Launch a local service
fiftyone delegated launch
List delegated operations¶
List delegated operations.
fiftyone delegated list [-h]
[-o OPERATOR]
[-d DATASET]
[-s STATE]
[--sort-by SORT_BY]
[--reverse]
[-l LIMIT]
Arguments
optional arguments:
-h, --help show this help message and exit
-o OPERATOR, --operator OPERATOR
only list operations for this operator
-d DATASET, --dataset DATASET
only list operations for this dataset
-s STATE, --state STATE
only list operations with this state. Supported
values are ('SCHEDULED', 'QUEUED', 'RUNNING', 'COMPLETED', 'FAILED')
--sort-by SORT_BY how to sort the operations. Supported values are
('SCHEDULED_AT', 'QUEUED_AT', 'STARTED_AT', COMPLETED_AT', 'FAILED_AT', 'OPERATOR')
--reverse whether to sort in reverse order
-l LIMIT, --limit LIMIT
a maximum number of operations to show
Examples
# List all delegated operations
fiftyone delegated list
# List some specific delegated operations
fiftyone delegated list \
--dataset quickstart \
--operator @voxel51/io/export_samples \
--state COMPLETED \
--sort-by COMPLETED_AT \
--limit 10
Delegated operation info¶
Prints information about a delegated operation.
fiftyone delegated info [-h] ID
Arguments
positional arguments:
ID the operation ID
optional arguments:
-h, --help show this help message and exit
Examples
# Print information about a delegated operation
fiftyone delegated info <id>
Mark delegated operations as failed¶
Manually mark delegated operations as failed.
fiftyone delegated fail [-h] [IDS ...]
Arguments
positional arguments:
IDS an operation ID or list of operation IDs
optional arguments:
-h, --help show this help message and exit
Examples
# Manually mark the specified operation(s) as FAILED
fiftyone delegated fail <id1> <id2> ...
Delete delegated operations¶
Delete delegated operations.
fiftyone delegated delete [-h] [IDS ...]
Arguments
positional arguments:
IDS an operation ID or list of operation IDs
optional arguments:
-h, --help show this help message and exit
Examples
# Delete the specified operation(s)
fiftyone delegated delete <id1> <id2> ...
Cleanup delegated operations¶
Cleanup delegated operations.
fiftyone delegated cleanup [-h]
[-o OPERATOR]
[-d DATASET]
[-s STATE]
[--orphan]
[--dry-run]
Arguments
optional arguments:
-h, --help show this help message and exit
-o OPERATOR, --operator OPERATOR
cleanup operations for this operator
-d DATASET, --dataset DATASET
cleanup operations for this dataset
-s STATE, --state STATE
delete operations in this state. Supported values
are ('QUEUED', 'COMPLETED', 'FAILED')
--orphan delete all operations associated with non-existent
datasets
--dry-run whether to print information rather than actually
deleting operations
Examples
# Delete all failed operations associated with a given dataset
fiftyone delegated cleanup --dataset quickstart --state FAILED
# Delete all delegated operations associated with non-existent datasets
fiftyone delegated cleanup --orphan
# Print information about operations rather than actually deleting them
fiftyone delegated cleanup --orphan --dry-run
FiftyOne plugins¶
Tools for working with FiftyOne plugins.
fiftyone plugins [-h] [--all-help] {list,info,download,requirements,create,enable,disable,delete} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{list,info,download,requirements,create,enable,disable,delete}
list List plugins that you've downloaded or created locally.
info Prints information about plugins that you've downloaded or created
download Download plugins from the web.
requirements Handles package requirements for plugins.
create Creates or initializes a plugin.
enable Enables the given plugin(s).
disable Disables the given plugin(s).
delete Delete plugins from your local machine.
List plugins¶
List plugins that you’ve downloaded or created locally.
fiftyone plugins list [-h] [-e] [-d] [-n]
Arguments
optional arguments:
-h, --help show this help message and exit
-e, --enabled only show enabled plugins
-d, --disabled only show disabled plugins
-n, --names-only only show plugin names
Examples
# List all locally available plugins
fiftyone plugins list
# List enabled plugins
fiftyone plugins list --enabled
# List disabled plugins
fiftyone plugins list --disabled
Plugin info¶
List plugins that you’ve downloaded or created locally.
fiftyone plugins info [-h] NAME
Arguments
positional arguments:
NAME the plugin name
optional arguments:
-h, --help show this help message and exit
Examples
# Prints information about a plugin
fiftyone plugins info <name>
Download plugins¶
Download plugins from the web.
When downloading plugins from GitHub, you can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone plugins download [-h] [-n [PLUGIN_NAMES ...]] [-o] URL_OR_GH_REPO
Arguments
positional arguments:
URL_OR_GH_REPO A URL or <user>/<repo>[/<ref>] of a GitHub repository
optional arguments:
-h, --help show this help message and exit
-n [PLUGIN_NAMES ...], --plugin-names [PLUGIN_NAMES ...]
a plugin name or list of plugin names to download
-o, --overwrite whether to overwrite existing plugins
Examples
# Download plugins from a GitHub repository URL
fiftyone plugins download <github-repo-url>
# Download plugins by specifying the GitHub repository details
fiftyone plugins download <user>/<repo>[/<ref>]
# Download specific plugins from a URL
fiftyone plugins download <url> --plugin-names <name1> <name2> <name3>
Plugin requirements¶
Handles package requirements for plugins.
fiftyone plugins requirements [-h] [-p] [-i] [-e] [--error-level LEVEL] NAME
Arguments
positional arguments:
NAME the plugin name
optional arguments:
-h, --help show this help message and exit
-p, --print print the requirements for the plugin
-i, --install install any requirements for the plugin
-e, --ensure ensure the requirements for the plugin are satisfied
--error-level LEVEL the error level (0=error, 1=warn, 2=ignore) to use when installing or ensuring plugin requirements
Examples
# Print requirements for a plugin
fiftyone plugins requirements <name> --print
# Install any requirements for the plugin
fiftyone plugins requirements <name> --install
# Ensures that the requirements for the plugin are satisfied
fiftyone plugins requirements <name> --ensure
Create plugins¶
Creates or initializes a plugin.
fiftyone plugins create [-h]
[-f [FILES ...]]
[-d OUTDIR]
[--label LABEL]
[--description DESCRIPTION]
[--version VERSION]
[-o]
[--kwargs KEY=VAL [KEY=VAL ...]]
[NAME ...]
Arguments
positional arguments:
NAME the plugin name(s)
optional arguments:
-h, --help show this help message and exit
-f [FILES ...], --from-files [FILES ...]
a directory or list of explicit filepaths to include in the plugin
-d OUTDIR, --outdir OUTDIR
a directory in which to create the plugin
--label LABEL a display name for the plugin
--description DESCRIPTION
a description for the plugin
--version VERSION an optional FiftyOne version requirement for the plugin
-o, --overwrite whether to overwrite existing plugins
--kwargs KEY=VAL [KEY=VAL ...]
additional keyword arguments to include in the plugin definition
Examples
# Initialize a new plugin
fiftyone plugins create <name>
# Create a plugin from existing files
fiftyone plugins create \
<name> \
--from-files /path/to/dir \
--label <label> \
--description <description>
Enable plugins¶
Enables the given plugin(s).
fiftyone plugins enable [-h] [-a] [NAME ...]
Arguments
positional arguments:
NAME the plugin name(s)
optional arguments:
-h, --help show this help message and exit
-a, --all whether to enable all plugins
Examples
# Enable a plugin
fiftyone plugins enable <name>
# Enable multiple plugins
fiftyone plugins enable <name1> <name2> ...
# Enable all plugins
fiftyone plugins enable --all
Disable plugins¶
Disables the given plugin(s).
fiftyone plugins disable [-h] [-a] [NAME ...]
Arguments
positional arguments:
NAME the plugin name(s)
optional arguments:
-h, --help show this help message and exit
-a, --all whether to disable all plugins
Examples
# Disable a plugin
fiftyone plugins disable <name>
# Disable multiple plugins
fiftyone plugins disable <name1> <name2> ...
# Disable all plugins
fiftyone plugins disable --all
Delete plugins¶
Delete plugins from your local machine.
fiftyone plugins delete [-h] [-a] [NAME ...]
Arguments
positional arguments:
NAME the plugin name(s)
optional arguments:
-h, --help show this help message and exit
-a, --all whether to delete all plugins
Examples
# Delete a plugin from local disk
fiftyone plugins delete <name>
# Delete multiple plugins from local disk
fiftyone plugins delete <name1> <name2> ...
# Delete all plugins from local disk
fiftyone plugins delete --all
FiftyOne utilities¶
FiftyOne utilities.
fiftyone utils [-h] [--all-help]
{compute-metadata,transform-images,transform-videos} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{compute-metadata,transform-images,transform-videos}
compute-metadata Populates the `metadata` field of all samples in the dataset.
transform-images Transforms the images in a dataset per the specified parameters.
transform-videos Transforms the videos in a dataset per the specified parameters.
Compute metadata¶
Populates the metadata field of all samples in the dataset.
fiftyone utils compute-metadata [-h] [-o] [-n NUM_WORKERS] [-s] DATASET_NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
-o, --overwrite whether to overwrite existing metadata
-n NUM_WORKERS, --num-workers NUM_WORKERS
a suggested number of worker processes to use
-s, --skip-failures whether to gracefully continue without raising an
error if metadata cannot be computed for a sample
Examples
# Populate all missing `metadata` sample fields
fiftyone utils compute-metadata <dataset-name>
# (Re)-populate the `metadata` field for all samples
fiftyone utils compute-metadata <dataset-name> --overwrite
Transform images¶
Transforms the images in a dataset per the specified parameters.
fiftyone utils transform-images [-h] [--size SIZE] [--min-size MIN_SIZE]
[--max-size MAX_SIZE] [-i INTERPOLATION]
[-e EXT] [-f] [--media-field MEDIA_FIELD]
[--output-field OUTPUT_FIELD]
[-o OUTPUT_DIR] [-r REL_DIR]
[--no-update-filepaths]
[-d] [-n NUM_WORKERS] [-s]
DATASET_NAME
Arguments
positional arguments:
DATASET_NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
--size SIZE a `width,height` for each image. A dimension can be
-1 if no constraint should be applied
--min-size MIN_SIZE a minimum `width,height` for each image. A dimension
can be -1 if no constraint should be applied
--max-size MAX_SIZE a maximum `width,height` for each image. A dimension
can be -1 if no constraint should be applied
-i INTERPOLATION, --interpolation INTERPOLATION
an optional `interpolation` argument for `cv2.resize()`
-e EXT, --ext EXT an image format to convert to (e.g., '.png' or '.jpg')
-f, --force-reencode whether to re-encode images whose parameters already
meet the specified values
--media-field MEDIA_FIELD
the input field containing the image paths to
transform
--output-field OUTPUT_FIELD
an optional field in which to store the paths to
the transformed images. By default, `media_field`
is updated in-place
-o OUTPUT_DIR, --output-dir OUTPUT_DIR
an optional output directory in which to write the
transformed images. If none is provided, the images
are updated in-place
-r REL_DIR, --rel-dir REL_DIR
an optional relative directory to strip from each
input filepath to generate a unique identifier that
is joined with `output_dir` to generate an output
path for each image
--no-update-filepaths
whether to store the output filepaths on the sample
collection
-d, --delete-originals
whether to delete the original images after transforming
-n NUM_WORKERS, --num-workers NUM_WORKERS
a suggested number of worker processes to use
-s, --skip-failures whether to gracefully continue without raising an
error if an image cannot be transformed
Examples
# Convert the images in the dataset to PNGs
fiftyone utils transform-images <dataset-name> --ext .png --delete-originals
# Ensure that no images in the dataset exceed 1920 x 1080
fiftyone utils transform-images <dataset-name> --max-size 1920,1080
Transform videos¶
Transforms the videos in a dataset per the specified parameters.
fiftyone utils transform-videos [-h] [--fps FPS] [--min-fps MIN_FPS]
[--max-fps MAX_FPS] [--size SIZE]
[--min-size MIN_SIZE] [--max-size MAX_SIZE]
[-r] [-f]
[--media-field MEDIA_FIELD]
[--output-field OUTPUT_FIELD]
[--output-dir OUTPUT_DIR]
[--rel-dir REL_DIR]
[--no-update-filepaths]
[-d] [-s] [-v]
DATASET_NAME
Arguments
positional arguments:
DATASET_NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
--fps FPS a frame rate at which to resample the videos
--min-fps MIN_FPS a minimum frame rate. Videos with frame rate below
this value are upsampled
--max-fps MAX_FPS a maximum frame rate. Videos with frame rate exceeding
this value are downsampled
--size SIZE a `width,height` for each frame. A dimension can be -1
if no constraint should be applied
--min-size MIN_SIZE a minimum `width,height` for each frame. A dimension
can be -1 if no constraint should be applied
--max-size MAX_SIZE a maximum `width,height` for each frame. A dimension
can be -1 if no constraint should be applied
-r, --reencode whether to re-encode the videos as H.264 MP4s
-f, --force-reencode whether to re-encode videos whose parameters already
meet the specified values
--media-field MEDIA_FIELD
the input field containing the video paths to
transform
--output-field OUTPUT_FIELD
an optional field in which to store the paths to
the transformed videos. By default, `media_field`
is updated in-place
--output-dir OUTPUT_DIR
an optional output directory in which to write the
transformed videos. If none is provided, the videos
are updated in-place
--rel-dir REL_DIR an optional relative directory to strip from each
input filepath to generate a unique identifier that
is joined with `output_dir` to generate an output
path for each video
--no-update-filepaths
whether to store the output filepaths on the sample
collection
-d, --delete-originals
whether to delete the original videos after transforming
-s, --skip-failures whether to gracefully continue without raising an
error if a video cannot be transformed
-v, --verbose whether to log the `ffmpeg` commands that are executed
Examples
# Re-encode the videos in the dataset as H.264 MP4s
fiftyone utils transform-videos <dataset-name> --reencode
# Ensure that no videos in the dataset exceed 1920 x 1080 and 30fps
fiftyone utils transform-videos <dataset-name> \
--max-size 1920,1080 --max-fps 30.0
FiftyOne Annotation¶
Tools for working with the FiftyOne annotation API.
fiftyone annotation [-h] [--all-help] {config} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{config}
config Tools for working with your FiftyOne annotation config.
Annotation Config¶
Tools for working with your FiftyOne annotation config.
fiftyone annotation config [-h] [-l] [FIELD]
Arguments
positional arguments:
FIELD an annotation config field to print
optional arguments:
-h, --help show this help message and exit
-l, --locate print the location of your annotation config on disk
Examples
# Print your entire annotation config
fiftyone annotation config
# Print a specific annotation config field
fiftyone annotation config <field>
# Print the location of your annotation config on disk (if one exists)
fiftyone annotation config --locate
FiftyOne App¶
Tools for working with the FiftyOne App.
fiftyone app [-h] [--all-help] {config,launch,view,connect} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{config,launch,view,connect}
config Tools for working with your App config.
launch Launch the FiftyOne App.
view View datasets in the App without persisting them to the database.
connect Connect to a remote FiftyOne App.
App Config¶
Tools for working with your FiftyOne App config.
fiftyone app config [-h] [-l] [FIELD]
Arguments
positional arguments:
FIELD an App config field to print
optional arguments:
-h, --help show this help message and exit
-l, --locate print the location of your App config on disk
Examples
# Print your entire App config
fiftyone app config
# Print a specific App config field
fiftyone app config <field>
# Print the location of your App config on disk (if one exists)
fiftyone app config --locate
Launch the App¶
Launch the FiftyOne App.
fiftyone app launch [-h] [-p PORT] [-A ADDRESS] [-b BROWSER] [-r] [-a] [-w WAIT] [NAME]
Arguments
positional arguments:
NAME the name of a dataset to open
optional arguments:
-h, --help show this help message and exit
-p PORT, --port PORT the port number to use
-A ADDRESS, --address ADDRESS
the address (server name) to use
-r, --remote whether to launch a remote App session
-b BROWSER, --browser BROWSER
the browser to use to open the App
-w WAIT, --wait WAIT the number of seconds to wait for a new App
connection before returning if all connections are
lost. If negative, the process will wait forever,
regardless of connections
Examples
# Launch the App
fiftyone app launch
# Launch the App with the given dataset loaded
fiftyone app launch <name>
# Launch a remote App session
fiftyone app launch ... --remote
# Launch an App session with a specific browser
fiftyone app launch ... --browser <name>
View datasets in App¶
View datasets in the FiftyOne App without persisting them to the database.
fiftyone app view [-h] [-n NAME] [-d DATASET_DIR] [-t TYPE] [-z NAME]
[-s SPLITS [SPLITS ...]] [--images-dir IMAGES_DIR]
[--images-patt IMAGES_PATT] [--videos-dir VIDEOS_DIR]
[--videos-patt VIDEOS_PATT] [-j JSON_PATH] [-p PORT]
[-A ADDRESS] [-r] [-a] [-w WAIT]
[-k KEY=VAL [KEY=VAL ...]]
Arguments
optional arguments:
-h, --help show this help message and exit
-n NAME, --name NAME a name for the dataset
-d DATASET_DIR, --dataset-dir DATASET_DIR
the directory containing the dataset to view
-t TYPE, --type TYPE the fiftyone.types.Dataset type of the dataset
-z NAME, --zoo-dataset NAME
the name of a zoo dataset to view
-s SPLITS [SPLITS ...], --splits SPLITS [SPLITS ...]
the dataset splits to load
--images-dir IMAGES_DIR
the path to a directory of images
--images-patt IMAGES_PATT
a glob pattern of images
--videos-dir VIDEOS_DIR
the path to a directory of videos
--videos-patt VIDEOS_PATT
a glob pattern of videos
-j JSON_PATH, --json-path JSON_PATH
the path to a samples JSON file to view
-p PORT, --port PORT the port number to use
-A ADDRESS, --address ADDRESS
the address (server name) to use
-r, --remote whether to launch a remote App session
-w WAIT, --wait WAIT the number of seconds to wait for a new App
connection before returning if all connections are
lost. If negative, the process will wait forever,
regardless of connections
-k KEY=VAL [KEY=VAL ...], --kwargs KEY=VAL [KEY=VAL ...]
additional type-specific keyword arguments for
`fiftyone.core.dataset.Dataset.from_dir()`
Examples
# View a dataset stored on disk in the App
fiftyone app view --dataset-dir <dataset-dir> --type <type>
# View a zoo dataset in the App
fiftyone app view --zoo-dataset <name> --splits <split1> ...
# View a directory of images in the App
fiftyone app view --images-dir <images-dir>
# View a glob pattern of images in the App
fiftyone app view --images-patt <images-patt>
# View a directory of videos in the App
fiftyone app view --videos-dir <videos-dir>
# View a glob pattern of videos in the App
fiftyone app view --videos-patt <videos-patt>
# View a dataset stored in JSON format on disk in the App
fiftyone app view --json-path <json-path>
# View the dataset in a remote App session
fiftyone app view ... --remote
# View a random subset of the data stored on disk in the App
fiftyone app view ... --kwargs max_samples=50 shuffle=True
Connect to remote App¶
Connect to a remote FiftyOne App in your web browser.
fiftyone app connect [-h] [-d DESTINATION] [-p PORT] [-A ADDRESS] [-l PORT]
[-i KEY]
Arguments
optional arguments:
-h, --help show this help message and exit
-d DESTINATION, --destination DESTINATION
the destination to connect to, e.g., [username@]hostname
-p PORT, --port PORT the remote port to connect to
-l PORT, --local-port PORT
the local port to use to serve the App
-i KEY, --ssh-key KEY
optional ssh key to use to login
Examples
# Connect to a remote App with port forwarding already configured
fiftyone app connect
# Connect to a remote App session
fiftyone app connect --destination <destination> --port <port>
# Connect to a remote App session using an ssh key
fiftyone app connect ... --ssh-key <path/to/key>
# Connect to a remote App using a custom local port
fiftyone app connect ... --local-port <port>
FiftyOne Brain¶
Tools for working with the FiftyOne Brain.
fiftyone brain [-h] [--all-help] {config} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{config}
config Tools for working with your FiftyOne Brain config.
Brain Config¶
Tools for working with your FiftyOne Brain config.
fiftyone brain config [-h] [-l] [FIELD]
Arguments
positional arguments:
FIELD a brain config field to print
optional arguments:
-h, --help show this help message and exit
-l, --locate print the location of your brain config on disk
Examples
# Print your entire brain config
fiftyone brain config
# Print a specific brain config field
fiftyone brain config <field>
# Print the location of your brain config on disk (if one exists)
fiftyone brain config --locate
FiftyOne Evaluation¶
Tools for working with the FiftyOne evaluation API.
fiftyone evaluation [-h] [--all-help] {config} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{config}
config Tools for working with your FiftyOne evaluation config.
Evaluation Config¶
Tools for working with your FiftyOne evaluation config.
fiftyone evaluation config [-h] [-l] [FIELD]
Arguments
positional arguments:
FIELD an evaluation config field to print
optional arguments:
-h, --help show this help message and exit
-l, --locate print the location of your evaluation config on disk
Examples
# Print your entire evaluation config
fiftyone evaluation config
# Print a specific evaluation config field
fiftyone evaluation config <field>
# Print the location of your evaluation config on disk (if one exists)
fiftyone evaluation config --locate
FiftyOne Zoo¶
Tools for working with the FiftyOne Zoo.
fiftyone zoo [-h] [--all-help] {datasets,models} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{datasets,models}
datasets Tools for working with the FiftyOne Dataset Zoo.
models Tools for working with the FiftyOne Model Zoo.
FiftyOne Dataset Zoo¶
Tools for working with the FiftyOne Dataset Zoo.
fiftyone zoo datasets [-h] [--all-help]
{list,find,info,download,load,delete} ...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{list,find,info,download,load,delete}
list List datasets in the FiftyOne Dataset Zoo.
find Locate a downloaded zoo dataset on disk.
info Print information about datasets in the FiftyOne Dataset Zoo.
download Download zoo datasets.
load Load zoo datasets as persistent FiftyOne datasets.
delete Deletes the local copy of the zoo dataset on disk.
List datasets in zoo¶
List datasets in the FiftyOne Dataset Zoo.
fiftyone zoo datasets list [-h] [-n] [-d] [-s SOURCE] [-t TAGS]
Arguments
optional arguments:
-h, --help show this help message and exit
-n, --names-only only show dataset names
-d, --downloaded-only
only show datasets that have been downloaded
-s SOURCE, --source SOURCE
only show datasets available from the specified source
-t TAGS, --tags TAGS only show datasets with the specified tag or list,of,tags
Examples
# List available datasets
fiftyone zoo datasets list
# List available dataset names
fiftyone zoo datasets list --names-only
# List downloaded datasets
fiftyone zoo datasets list --downloaded-only
# List available datasets from the given source
fiftyone zoo datasets list --source <source>
# List available datasets with the given tag
fiftyone zoo datasets list --tags <tag>
Find zoo datasets on disk¶
Locate a downloaded zoo dataset on disk.
fiftyone zoo datasets find [-h] [-s SPLIT] NAME_OR_URL
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the dataset
optional arguments:
-h, --help show this help message and exit
-s SPLIT, --split SPLIT
Examples
# Print the location of a downloaded zoo dataset on disk
fiftyone zoo datasets find <name>
# Print the location of a remotely-sourced zoo dataset on disk
fiftyone zoo datasets find https://github.com/<user>/<repo>
fiftyone zoo datasets find <url>
# Print the location of a specific split of a dataset
fiftyone zoo datasets find <name> --split <split>
Show zoo dataset info¶
Print information about datasets in the FiftyOne Dataset Zoo.
fiftyone zoo datasets info [-h] NAME_OR_URL
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the dataset
optional arguments:
-h, --help show this help message and exit
Examples
# Print information about a zoo dataset
fiftyone zoo datasets info <name>
# Print information about a remote zoo dataset
fiftyone zoo datasets info https://github.com/<user>/<repo>
fiftyone zoo datasets info <url>
Download zoo datasets¶
Download zoo datasets.
When downloading remotely-sourced zoo datasets, you can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]a publicly accessible URL of an archive (eg zip or tar) file
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone zoo datasets download [-h] [-s SPLITS [SPLITS ...]]
[-k KEY=VAL [KEY=VAL ...]]
NAME_OR_URL
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the dataset
optional arguments:
-h, --help show this help message and exit
-s SPLITS [SPLITS ...], --splits SPLITS [SPLITS ...]
the dataset splits to download
-k KEY=VAL [KEY=VAL ...], --kwargs KEY=VAL [KEY=VAL ...]
optional dataset-specific keyword arguments for
`fiftyone.zoo.download_zoo_dataset()`
Examples
# Download a zoo dataset
fiftyone zoo datasets download <name>
# Download a remotely-sourced zoo dataset
fiftyone zoo datasets download https://github.com/<user>/<repo>
fiftyone zoo datasets download <url>
# Download the specified split(s) of a zoo dataset
fiftyone zoo datasets download <name> --splits <split1> ...
# Download a zoo dataset that requires extra keyword arguments
fiftyone zoo datasets download <name> \
--kwargs source_dir=/path/to/source/files
Load zoo datasets¶
Load zoo datasets as persistent FiftyOne datasets.
When loading remotely-sourced zoo datasets, you can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]a publicly accessible URL of an archive (eg zip or tar) file
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone zoo datasets load [-h] [-s SPLITS [SPLITS ...]]
[-n DATASET_NAME] [-k KEY=VAL [KEY=VAL ...]]
NAME_OR_URL
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the dataset
optional arguments:
-h, --help show this help message and exit
-s SPLITS [SPLITS ...], --splits SPLITS [SPLITS ...]
the dataset splits to load
-n DATASET_NAME, --dataset-name DATASET_NAME
a custom name to give the FiftyOne dataset
-k KEY=VAL [KEY=VAL ...], --kwargs KEY=VAL [KEY=VAL ...]
additional dataset-specific keyword arguments for
`fiftyone.zoo.load_zoo_dataset()`
Examples
# Load the zoo dataset with the given name
fiftyone zoo datasets load <name>
# Load a remotely-sourced zoo dataset
fiftyone zoo datasets load https://github.com/<user>/<repo>
fiftyone zoo datasets load <url>
# Load the specified split(s) of a zoo dataset
fiftyone zoo datasets load <name> --splits <split1> ...
# Load a zoo dataset with a custom name
fiftyone zoo datasets load <name> --dataset-name <dataset-name>
# Load a zoo dataset that requires custom keyword arguments
fiftyone zoo datasets load <name> \
--kwargs source_dir=/path/to/source_files
# Load a random subset of a zoo dataset
fiftyone zoo datasets load <name> \
--kwargs max_samples=50 shuffle=True
Delete zoo datasets¶
Deletes the local copy of the zoo dataset on disk.
fiftyone zoo datasets delete [-h] [-s SPLIT] NAME
Arguments
positional arguments:
NAME the name of the dataset
optional arguments:
-h, --help show this help message and exit
-s SPLIT, --split SPLIT
a dataset split
Examples
# Delete an entire zoo dataset from disk
fiftyone zoo datasets delete <name>
# Delete a specific split of a zoo dataset from disk
fiftyone zoo datasets delete <name> --split <split>
FiftyOne Model Zoo¶
Tools for working with the FiftyOne Model Zoo.
fiftyone zoo models [-h] [--all-help]
{list,find,info,requirements,download,apply,embed,delete,list-sources,register-source,delete-source}
...
Arguments
optional arguments:
-h, --help show this help message and exit
--all-help show help recursively and exit
available commands:
{list,find,info,requirements,download,apply,embed,delete,register-source,delete-source}
list List models in the FiftyOne Model Zoo.
find Locate the downloaded zoo model on disk.
info Print information about models in the FiftyOne Model Zoo.
requirements Handles package requirements for zoo models.
download Download zoo models.
apply Apply zoo models to datasets.
embed Generate embeddings for datasets with zoo models.
delete Deletes the local copy of the zoo model on disk.
list-sources Lists remote zoo model sources that are registered locally.
register-source Registers a remote source of zoo models.
delete-source Deletes the remote source and all downloaded models associated with it.
List models in zoo¶
List models in the FiftyOne Model Zoo.
fiftyone zoo models list [-h] [-n] [-d] [-t TAGS] [-s SOURCE]
Arguments
optional arguments:
-h, --help show this help message and exit
-n, --names-only only show model names
-d, --downloaded-only
only show models that have been downloaded
-t TAGS, --tags TAGS only show models with the specified tag or list,of,tags
-s SOURCE, --source SOURCE
only show models available from the specified remote source
Examples
# List available models
fiftyone zoo models list
# List available models (names only)
fiftyone zoo models list --names-only
# List downloaded models
fiftyone zoo models list --downloaded-only
# List available models with the given tag
fiftyone zoo models list --tags <tag>
# List available models from the given remote source
fiftyone zoo models list --source <source>
Find zoo models on disk¶
Locate the downloaded zoo model on disk.
fiftyone zoo models find [-h] NAME
Arguments
positional arguments:
NAME the name of the model
optional arguments:
-h, --help show this help message and exit
Examples
# Print the location of the downloaded zoo model on disk
fiftyone zoo models find <name>
Show zoo model info¶
Print information about models in the FiftyOne Model Zoo.
fiftyone zoo models info [-h] NAME
Arguments
positional arguments:
NAME the name of the model
optional arguments:
-h, --help show this help message and exit
Examples
# Print information about a zoo model
fiftyone zoo models info <name>
Zoo model requirements¶
Handles package requirements for zoo models.
fiftyone zoo models requirements [-h] [-p] [-i] [-e]
[--error-level LEVEL]
NAME
Arguments
positional arguments:
NAME the name of the model
optional arguments:
-h, --help show this help message and exit
-p, --print print the requirements for the zoo model
-i, --install install any requirements for the zoo model
-e, --ensure ensure the requirements for the zoo model are satisfied
--error-level LEVEL the error level (0=error, 1=warn, 2=ignore) to use
when installing or ensuring model requirements
Examples
# Print requirements for a zoo model
fiftyone zoo models requirements <name> --print
# Install any requirements for the zoo model
fiftyone zoo models requirements <name> --install
# Ensures that the requirements for the zoo model are satisfied
fiftyone zoo models requirements <name> --ensure
Download zoo models¶
Download zoo models.
When downloading remotely-sourced zoo models, you can provide any of the following:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone zoo models download [-h] [-n MODEL_NAME] [-o] NAME_OR_URL
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the model
optional arguments:
-h, --help show this help message and exit
-n MODEL_NAME, --model-name MODEL_NAME
the specific model to download, if `name_or_url` is
a remote source
-o, --overwrite whether to overwrite any existing model files
Examples
# Download a zoo model
fiftyone zoo models download <name>
# Download a remotely-sourced zoo model
fiftyone zoo models download https://github.com/<user>/<repo> \
--model-name <name>
fiftyone zoo models download <url> --model-name <name>
Apply zoo models to datasets¶
Apply zoo models to datasets.
When applying remotely-sourced zoo models, you can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]a publicly accessible URL of an archive (eg zip or tar) file
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone zoo models apply [-h] [-n MODEL_NAME] [-b BATCH_SIZE] [-t THRESH]
[-l] [-i] [--error-level LEVEL]
NAME_OR_URL DATASET_NAME LABEL_FIELD
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the zoo model
DATASET_NAME the name of the FiftyOne dataset to process
LABEL_FIELD the name of the field in which to store the predictions
optional arguments:
-h, --help show this help message and exit
-n MODEL_NAME, --model-name MODEL_NAME
the specific model to apply, if `name_or_url` is a
remote source
-b BATCH_SIZE, --batch-size BATCH_SIZE
an optional batch size to use during inference
-t THRESH, --confidence-thresh THRESH
an optional confidence threshold to apply to any
applicable labels generated by the model
-l, --store-logits store logits for the predictions
-i, --install install any requirements for the zoo model
--error-level LEVEL the error level (0=error, 1=warn, 2=ignore) to use
when installing or ensuring model requirements
Examples
# Apply a zoo model to a dataset
fiftyone zoo models apply <model-name> <dataset-name> <label-field>
# Apply a remotely-sourced zoo model to a dataset
fiftyone zoo models apply https://github.com/<user>/<repo> \
<dataset-name> <label-field> --model-name <model-name>
fiftyone zoo models apply <url> \
<dataset-name> <label-field> --model-name <model-name>
# Apply a zoo classifier with some customized parameters
fiftyone zoo models apply \
<model-name> <dataset-name> <label-field> \
--confidence-thresh 0.7 \
--store-logits \
--batch-size 32
Generate embeddings with zoo models¶
Generate embeddings for datasets with zoo models.
When applying remotely-sourced zoo models, you can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]a publicly accessible URL of an archive (eg zip or tar) file
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone zoo models embed [-h] [-n MODEL_NAME] [-b BATCH_SIZE] [-i]
[--error-level LEVEL]
NAME_OR_URL DATASET_NAME EMBEDDINGS_FIELD
Arguments
positional arguments:
NAME_OR_URL the name or remote location of the zoo model
DATASET_NAME the name of the FiftyOne dataset to process
EMBEDDINGS_FIELD the name of the field in which to store the embeddings
optional arguments:
-h, --help show this help message and exit
-n MODEL_NAME, --model-name MODEL_NAME
the specific model to apply, if `name_or_url` is a
remote source
-b BATCH_SIZE, --batch-size BATCH_SIZE
an optional batch size to use during inference
-i, --install install any requirements for the zoo model
--error-level LEVEL the error level (0=error, 1=warn, 2=ignore) to use
when installing or ensuring model requirements
Examples
# Generate embeddings for a dataset with a zoo model
fiftyone zoo models embed <model-name> <dataset-name> <embeddings-field>
# Generate embeddings for a dataset with a remotely-sourced zoo model
fiftyone zoo models embed https://github.com/<user>/<repo> \
<dataset-name> <embeddings-field> --model-name <model-name>
fiftyone zoo models embed <url> \
<dataset-name> <embeddings-field> --model-name <model-name>
Delete zoo models¶
Deletes the local copy of the zoo model on disk.
fiftyone zoo models delete [-h] NAME
Arguments
positional arguments:
NAME the name of the model
optional arguments:
-h, --help show this help message and exit
Examples
# Delete the zoo model from disk
fiftyone zoo models delete <name>
List zoo model sources¶
Lists remote zoo model sources that are registered locally.
fiftyone zoo models list-sources [-h]
Examples
# Lists the registered remote zoo model sources
fiftyone zoo models list-sources
Register zoo model sources¶
Registers a remote source of zoo models.
You can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]a publicly accessible URL of an archive (eg zip or tar) file
Note
To download from a private GitHub repository that you have access to,
provide your GitHub personal access token by setting the GITHUB_TOKEN
environment variable.
fiftyone zoo models register-source [-h] [-o] URL_OR_GH_REPO
Arguments
positional arguments:
URL_OR_GH_REPO the remote source to register
optional arguments:
-h, --help show this help message and exit
-o, --overwrite whether to overwrite any existing files
Examples
# Register a remote zoo model source
fiftyone zoo models register-source https://github.com/<user>/<repo>
fiftyone zoo models register-source <url>
Delete zoo model sources¶
Deletes the remote source and all downloaded models associated with it.
You can provide any of the following formats:
a GitHub repo URL like
https://github.com/<user>/<repo>a GitHub ref like
https://github.com/<user>/<repo>/tree/<branch>orhttps://github.com/<user>/<repo>/commit/<commit>a GitHub ref string like
<user>/<repo>[/<ref>]a publicly accessible URL of an archive (eg zip or tar) file
fiftyone zoo models delete-source [-h] URL_OR_GH_REPO
Arguments
positional arguments:
URL_OR_GH_REPO the remote source to delete
optional arguments:
-h, --help show this help message and exit
Examples
# Delete a remote zoo model source
fiftyone zoo models delete-source https://github.com/<user>/<repo>
fiftyone zoo models delete-source <url>