Documentation
Search…
wandb.Artifact
Flexible and lightweight building block for dataset and model versioning.
1
Artifact(
2
name: str,
3
type: str,
4
description: Optional[str] = None,
5
metadata: Optional[dict] = None,
6
incremental: Optional[bool] = None,
7
use_as: Optional[str] = None
8
) -> None
Copied!
Constructs an empty artifact whose contents can be populated using its add family of functions. Once the artifact has all the desired files, you can call wandb.log_artifact() to log it.
Arguments
Text
name
(str) A human-readable name for this artifact, which is how you can identify this artifact in the UI or reference it in use_artifact calls. Names can contain letters, numbers, underscores, hyphens, and dots. The name must be unique across a project.
type
(str) The type of the artifact, which is used to organize and differentiate artifacts. Common types include dataset or model, but you can use any string containing letters, numbers, underscores, hyphens, and dots.
description
(str, optional) Free text that offers a description of the artifact. The description is markdown rendered in the UI, so this is a good place to place tables, links, etc.
metadata
(dict, optional) Structured data associated with the artifact, for example class distribution of a dataset. This will eventually be queryable and plottable in the UI. There is a hard limit of 100 total keys.

Examples:

Basic usage
1
wandb.init()
2
3
artifact = wandb.Artifact('mnist', type='dataset')
4
artifact.add_dir('mnist/')
5
wandb.log_artifact(artifact)
Copied!
Raises
Text
Exception
if problem.
Returns
Text
An Artifact object.
Attributes
Text
aliases
Returns: (list): A list of the aliases associated with this artifact. The list is mutable and calling save() will persist all alias changes.
commit_hash
Returns: (str): The artifact's commit hash which is used in http URLs
description
Returns: (str): Free text that offers a description of the artifact. The description is markdown rendered in the UI, so this is a good place to put links, etc.
digest
Returns: (str): The artifact's logical digest, a checksum of its contents. If an artifact has the same digest as the current latest version, then log_artifact is a no-op.
entity
Returns: (str): The name of the entity this artifact belongs to.
id
Returns: (str): The artifact's ID
manifest
Returns: (ArtifactManifest): The artifact's manifest, listing all of its contents. You cannot add more files to an artifact once you've retrieved its manifest.
metadata
Returns: (dict): Structured data associated with the artifact, for example class distribution of a dataset. This will eventually be queryable and plottable in the UI. There is a hard limit of 100 total keys.
name
Returns: (str): The artifact's name
project
Returns: (str): The name of the project this artifact belongs to.
size
Returns: (int): The size in bytes of the artifact. Includes any references tracked by this artifact.
state
Returns: (str): The state of the artifact, which can be one of "PENDING", "COMMITTED", or "DELETED".
type
Returns: (str): The artifact's type
version
Returns: (int): The version of this artifact. For example, if this is the first version of an artifact, its version will be 'v0'.

Methods

add

1
add(
2
obj: data_types.WBValue,
3
name: str
4
) -> ArtifactEntry
Copied!
Adds wandb.WBValue obj to the artifact.
1
obj = artifact.get(name)
Copied!
Arguments
Text
obj
(wandb.WBValue) The object to add. Currently support one of Bokeh, JoinedTable, PartitionedTable, Table, Classes, ImageMask, BoundingBoxes2D, Audio, Image, Video, Html, Object3D
name
(str) The path within the artifact to add the object.
Returns
Text
ArtifactManifestEntry
the added manifest entry

Examples:

Basic usage
1
artifact = wandb.Artifact('my_table', 'dataset')
2
table = wandb.Table(columns=["a", "b", "c"], data=[[i, i*2, 2**i]])
3
artifact.add(table, "my_table")
4
5
wandb.log_artifact(artifact)
Copied!
Retrieving an object:
1
artifact = wandb.use_artifact('my_table:latest')
2
table = artifact.get("my_table")
Copied!

add_dir

1
add_dir(
2
local_path: str,
3
name: Optional[str] = None
4
) -> None
Copied!
Adds a local directory to the artifact.
Arguments
Text
local_path
(str) The path to the directory being added.
name
(str, optional) The path within the artifact to use for the directory being added. Defaults to files being added under the root of the artifact.

Examples:

Adding a directory without an explicit name:
1
artifact.add_dir('my_dir/') # All files in `my_dir/` are added at the root of the artifact.
Copied!
Adding a directory without an explicit name:
1
artifact.add_dir('my_dir/', path='destination') # All files in `my_dir/` are added under `destination/`.
Copied!
Raises
Text
Exception
if problem.
Returns
Text
None

add_file

1
add_file(
2
local_path: str,
3
name: Optional[str] = None,
4
is_tmp: Optional[bool] = (False)
5
) -> ArtifactEntry
Copied!
Adds a local file to the artifact.
Arguments
Text
local_path
(str) The path to the file being added.
name
(str, optional) The path within the artifact to use for the file being added. Defaults to the basename of the file.
is_tmp
(bool, optional) If true, then the file is renamed deterministically to avoid collisions. (default: False)

Examples:

Adding a file without an explicit name:
1
artifact.add_file('path/to/file.txt') # Added as `file.txt'
Copied!
Adding a file with an explicit name:
1
artifact.add_file('path/to/file.txt', name='new/path/file.txt') # Added as 'new/path/file.txt'
Copied!
Raises
Text
Exception
if problem
Returns
Text
ArtifactManifestEntry
the added manifest entry

add_reference

1
add_reference(
2
uri: Union[ArtifactEntry, str],
3
name: Optional[str] = None,
4
checksum: bool = (True),
5
max_objects: Optional[int] = None
6
) -> Sequence[ArtifactEntry]
Copied!
Adds a reference denoted by a URI to the artifact. Unlike adding files or directories, references are NOT uploaded to W&B. However, artifact methods such as download() can be used regardless of whether the artifact contains references or uploaded files.
By default, W&B offers special handling for the following schemes:
    http(s): The size and digest of the file will be inferred by the Content-Length and the ETag response headers returned by the server.
    s3: The checksum and size will be pulled from the object metadata. If bucket versioning is enabled, then the version ID is also tracked.
    gs: The checksum and size will be pulled from the object metadata. If bucket versioning is enabled, then the version ID is also tracked.
    file: The checksum and size will be pulled from the file system. This scheme is useful if you have an NFS share or other externally mounted volume containing files you wish to track but not necessarily upload.
For any other scheme, the digest is just a hash of the URI and the size is left blank.
Arguments
Text
uri
(str) The URI path of the reference to add. Can be an object returned from Artifact.get_path to store a reference to another artifact's entry.
name
(str) The path within the artifact to place the contents of this reference
checksum
(bool, optional) Whether or not to checksum the resource(s) located at the reference URI. Checksumming is strongly recommended as it enables automatic integrity validation, however it can be disabled to speed up artifact creation. (default: True)
max_objects
(int, optional) The maximum number of objects to consider when adding a reference that points to directory or bucket store prefix. For S3 and GCS, this limit is 10,000 by default but is uncapped for other URI schemes. (default: None)
Raises
Text
Exception
If problem.
Returns
Text
List[ArtifactManifestEntry]: The added manifest entries.

Examples:

Adding an HTTP link:
1
# Adds `file.txt` to the root of the artifact as a reference
2
artifact.add_reference('http://myserver.com/file.txt')
Copied!
Adding an S3 prefix without an explicit name:
1
# All objects under `prefix/` will be added at the root of the artifact.
2
artifact.add_reference('s3://mybucket/prefix')
Copied!
Adding a GCS prefix with an explicit name:
1
# All objects under `prefix/` will be added under `path/` at the top of the artifact.
2
artifact.add_reference('gs://mybucket/prefix', name='path')
Copied!

checkout

1
checkout(
2
root: Optional[str] = None
3
) -> str
Copied!
Replaces the specified root directory with the contents of the artifact.
WARNING: This will DELETE all files in root that are not included in the artifact.
Arguments
Text
root
(str, optional) The directory to replace with this artifact's files.
Returns
Text
(str): The path to the checked out contents.

delete

1
delete() -> None
Copied!
Deletes this artifact, cleaning up all files associated with it.
NOTE: Deletion is permanent and CANNOT be undone.
Returns
Text
None

download

1
download(
2
root: str = None,
3
recursive: bool = (False)
4
) -> str
Copied!
Downloads the contents of the artifact to the specified root directory.
NOTE: Any existing files at root are left untouched. Explicitly delete root before calling download if you want the contents of root to exactly match the artifact.
Arguments
Text
root
(str, optional) The directory in which to download this artifact's files.
recursive
(bool, optional) If true, then all dependent artifacts are eagerly downloaded. Otherwise, the dependent artifacts are downloaded as needed.
Returns
Text
(str): The path to the downloaded contents.

finalize

1
finalize() -> None
Copied!
Marks this artifact as final, which disallows further additions to the artifact. This happens automatically when calling log_artifact.
Returns
Text
None

get

1
get(
2
name: str
3
) -> data_types.WBValue
Copied!
Gets the WBValue object located at the artifact relative name.
NOTE: This will raise an error unless the artifact has been fetched using use_artifact, fetched using the API, or wait() has been called.
Arguments
Text
name
(str) The artifact relative name to get
Raises
Text
Exception
if problem

Examples:

Basic usage
1
# Run logging the artifact
2
with wandb.init() as r:
3
artifact = wandb.Artifact('my_dataset', type='dataset')
4
table = wandb.Table(columns=["a", "b", "c"], data=[[i, i*2, 2**i]])
5
artifact.add(table, "my_table")
6
wandb.log_artifact(artifact)
7
8
# Run using the artifact
9
with wandb.init() as r:
10
artifact = r.use_artifact('my_dataset:latest')
11
table = r.get('my_table')
Copied!

get_added_local_path_name

1
get_added_local_path_name(
2
local_path: str
3
) -> Optional[str]
Copied!
Get the artifact relative name of a file added by a local filesystem path.
Arguments
Text
local_path
(str) The local path to resolve into an artifact relative name.
Returns
Text
str
The artifact relative name.

Examples:

Basic usage
1
artifact = wandb.Artifact('my_dataset', type='dataset')
2
artifact.add_file('path/to/file.txt', name='artifact/path/file.txt')
3
4
# Returns `artifact/path/file.txt`:
5
name = artifact.get_added_local_path_name('path/to/file.txt')
Copied!

get_path

1
get_path(
2
name: str
3
) -> ArtifactEntry
Copied!
Gets the path to the file located at the artifact relative name.
NOTE: This will raise an error unless the artifact has been fetched using use_artifact, fetched using the API, or wait() has been called.
Arguments
Text
name
(str) The artifact relative name to get
Raises
Text
Exception
if problem

Examples:

Basic usage
1
# Run logging the artifact
2
with wandb.init() as r:
3
artifact = wandb.Artifact('my_dataset', type='dataset')
4
artifact.add_file('path/to/file.txt')
5
wandb.log_artifact(artifact)
6
7
# Run using the artifact
8
with wandb.init() as r:
9
artifact = r.use_artifact('my_dataset:latest')
10
path = artifact.get_path('file.txt')
11
12
# Can now download 'file.txt' directly:
13
path.download()
Copied!

json_encode

1
json_encode() -> Dict[str, Any]
Copied!

logged_by

1
logged_by() -> "wandb.apis.public.Run"
Copied!
Returns: (Run): The run that first logged this artifact.

new_file

1
@contextlib.contextmanager
2
new_file(
3
name: str,
4
mode: str = "w"
5
) -> Generator[IO, None, None]
Copied!
Open a new temporary file that will be automatically added to the artifact.
Arguments
Text
name
(str) The name of the new file being added to the artifact.
mode
(str, optional) The mode in which to open the new file.

Examples:

1
artifact = wandb.Artifact('my_data', type='dataset')
2
with artifact.new_file('hello.txt') as f:
3
f.write('hello!')
4
wandb.log_artifact(artifact)
Copied!
Returns
Text
(file): A new file object that can be written to. Upon closing, the file will be automatically added to the artifact.

save

1
save(
2
project: Optional[str] = None,
3
settings: Optional['wandb.wandb_sdk.wandb_settings.Settings'] = None
4
) -> None
Copied!
Persists any changes made to the artifact. If currently in a run, that run will log this artifact. If not currently in a run, a run of type "auto" will be created to track this artifact.
Arguments
Text
project
(str, optional) A project to use for the artifact in the case that a run is not already in context
settings
(wandb.Settings, optional) A settings object to use when initializing an automatic run. Most commonly used in testing harness.
Returns
Text
None

used_by

1
used_by() -> List['wandb.apis.public.Run']
Copied!
Returns: (list): A list of the runs that have used this artifact.

verify

1
verify(
2
root: Optional[str] = None
3
) -> bool
Copied!
Verify that the actual contents of an artifact at a specified directory root match the expected contents of the artifact according to its manifest.
All files in the directory are checksummed and the checksums are then cross-referenced against the artifact's manifest.
NOTE: References are not verified.
Arguments
Text
root
(str, optional) The directory to verify. If None artifact will be downloaded to './artifacts/<self.name>/'
Raises
Text
(ValueError): If the verification fails.

wait

1
wait() -> ArtifactInterface
Copied!
Waits for this artifact to finish logging, if needed.
Returns
Text
Artifact

__getitem__

1
__getitem__(
2
name: str
3
) -> Optional[data_types.WBValue]
Copied!
Gets the WBValue object located at the artifact relative name.
NOTE: This will raise an error unless the artifact has been fetched using use_artifact, fetched using the API, or wait() has been called.
Arguments
Text
name
(str) The artifact relative name to get
Raises
Text
Exception
if problem

Examples:

Basic usage
1
artifact = wandb.Artifact('my_table', 'dataset')
2
table = wandb.Table(columns=["a", "b", "c"], data=[[i, i*2, 2**i]])
3
artifact["my_table"] = table
4
5
wandb.log_artifact(artifact)
Copied!
Retrieving an object:
1
artifact = wandb.use_artifact('my_table:latest')
2
table = artifact["my_table"]
Copied!
Last modified 9m ago