This is the multi-page printable view of this section. Click here to print.

Return to the regular view of this page.

Analytics and Query API

1: api
2: artifacts
3: files
4: history
5: jobs
6: projects
7: query_generator
8: reports
9: runs
10: sweeps
11: teams
12: users

Query and analyze data logged to W&B.

1 - api

View the source code

`module` `wandb.apis.public`

Use the Public API to export or update data that you have saved to W&B.

Before using this API, you’ll want to log data from your script — check the Quickstart for more details.

You might use the Public API to

update metadata or metrics for an experiment after it has been completed,
pull down your results as a dataframe for post-hoc analysis in a Jupyter notebook, or
check your saved model artifacts for those tagged as ready-to-deploy.

For more on using the Public API, check out our guide.

`class` `RetryingClient`

`method` `RetryingClient.init`

__init__(client: wandb_gql.client.Client)

`property` RetryingClient.app_url

`property` RetryingClient.server_info

`method` `RetryingClient.execute`

execute(*args, **kwargs)

`method` `RetryingClient.version_supported`

version_supported(min_version: str) → bool

`class` `Api`

Used for querying the wandb server.

Examples: Most common way to initialize wandb.Api()

Args:

overrides: (dict) You can set base_url if you are using a wandb server
other than https: //api.wandb.ai. You can also set defaults for entity, project, and run.

`method` `Api.init`

__init__(
    overrides: Optional[Dict[str, Any]] = None,
    timeout: Optional[int] = None,
    api_key: Optional[str] = None
) → None

`property` Api.api_key

`property` Api.client

`property` Api.default_entity

`property` Api.user_agent

`property` Api.viewer

`method` `Api.artifact`

artifact(name: str, type: Optional[str] = None)

Return a single artifact by parsing path in the form project/name or entity/project/name.

Args:

name: (str) An artifact name. May be prefixed with project/ or entity/project/. If no entity is specified in the name, the Run or API setting’s entity is used. Valid names can be in the following forms:
name: version
name: alias
type: (str, optional) The type of artifact to fetch.

Returns: An Artifact object.

Raises:

ValueError: If the artifact name is not specified.
ValueError: If the artifact type is specified but does not match the type of the fetched artifact.

Note:

This method is intended for external use only. Do not call api.artifact() within the wandb repository code.

`method` `Api.artifact_collection`

artifact_collection(type_name: str, name: str) → public.ArtifactCollection

Return a single artifact collection by type and parsing path in the form entity/project/name.

Args:

type_name: (str) The type of artifact collection to fetch.
name: (str) An artifact collection name. May be prefixed with entity/project.

Returns: An ArtifactCollection object.

`method` `Api.artifact_collection_exists`

artifact_collection_exists(name: str, type: str) → bool

Return whether an artifact collection exists within a specified project and entity.

Args:

name: (str) An artifact collection name. May be prefixed with entity/project. If entity or project is not specified, it will be inferred from the override params if populated. Otherwise, entity will be pulled from the user settings and project will default to “uncategorized”.
type: (str) The type of artifact collection

Returns: True if the artifact collection exists, False otherwise.

`method` `Api.artifact_collections`

artifact_collections(
    project_name: str,
    type_name: str,
    per_page: int = 50
) → public.ArtifactCollections

Return a collection of matching artifact collections.

Args:

project_name: (str) The name of the project to filter on.
type_name: (str) The name of the artifact type to filter on.
per_page: (int) Sets the page size for query pagination. Usually there is no reason to change this.

Returns: An iterable ArtifactCollections object.

`method` `Api.artifact_exists`

artifact_exists(name: str, type: Optional[str] = None) → bool

Return whether an artifact version exists within a specified project and entity.

Args:

name: (str) An artifact name. May be prefixed with entity/project. If entity or project is not specified, it will be inferred from the override params if populated. Otherwise, entity will be pulled from the user settings and project will default to “uncategorized”. Valid names can be in the following forms:
name: version
name: alias
type: (str, optional) The type of artifact

Returns: True if the artifact version exists, False otherwise.

`method` `Api.artifact_type`

artifact_type(
    type_name: str,
    project: Optional[str] = None
) → public.ArtifactType

Return the matching ArtifactType.

Args:

type_name: (str) The name of the artifact type to retrieve.
project: (str, optional) If given, a project name or path to filter on.

Returns: An ArtifactType object.

`method` `Api.artifact_types`

artifact_types(project: Optional[str] = None) → public.ArtifactTypes

Return a collection of matching artifact types.

Args:

project: (str, optional) If given, a project name or path to filter on.

Returns: An iterable ArtifactTypes object.

`method` `Api.artifact_versions`

artifact_versions(type_name, name, per_page=50)

Deprecated, use artifacts(type_name, name) instead.

`method` `Api.artifacts`

artifacts(
    type_name: str,
    name: str,
    per_page: int = 50,
    tags: Optional[List[str]] = None
) → public.Artifacts

Return an Artifacts collection from the given parameters.

Args:

type_name: (str) The type of artifacts to fetch.
name: (str) An artifact collection name. May be prefixed with entity/project.
per_page: (int) Sets the page size for query pagination. Usually there is no reason to change this.
tags: (list[str], optional) Only return artifacts with all of these tags.

Returns: An iterable Artifacts object.

`method` `Api.create_project`

create_project(name: str, entity: str) → None

Create a new project.

Args:

name: (str) The name of the new project.
entity: (str) The entity of the new project.

`method` `Api.create_run`

create_run(
    run_id: Optional[str] = None,
    project: Optional[str] = None,
    entity: Optional[str] = None
) → public.Run

Create a new run.

Args:

run_id: (str, optional) The ID to assign to the run, if given. The run ID is automatically generated by default, so in general, you do not need to specify this and should only do so at your own risk.
project: (str, optional) If given, the project of the new run.
entity: (str, optional) If given, the entity of the new run.

Returns: The newly created Run.

`method` `Api.create_run_queue`

create_run_queue(
    name: str,
    type: 'public.RunQueueResourceType',
    entity: Optional[str] = None,
    prioritization_mode: Optional[ForwardRef('public.RunQueuePrioritizationMode')] = None,
    config: Optional[dict] = None,
    template_variables: Optional[dict] = None
) → public.RunQueue

Create a new run queue (launch).

Args:

name: (str) Name of the queue to create
type: (str) Type of resource to be used for the queue. One of “local-container”, “local-process”, “kubernetes”, “sagemaker”, or “gcp-vertex”.
entity: (str) Optional name of the entity to create the queue. If None, will use the configured or default entity.
prioritization_mode: (str) Optional version of prioritization to use. Either “V0” or None
config: (dict) Optional default resource configuration to be used for the queue. Use handlebars (eg. {{var}}) to specify template variables.
template_variables: (dict) A dictionary of template variable schemas to be used with the config. Expected format of: `{
"var-name": {
"schema": {
"type": (“string”, “number”, or “integer”),
"default": (optional value),
"minimum": (optional minimum),
"maximum": (optional maximum),
"enum": […"(options)"] } } }`

Returns: The newly created RunQueue

Raises: ValueError if any of the parameters are invalid wandb.Error on wandb API errors

`method` `Api.create_team`

create_team(team, admin_username=None)

Create a new team.

Args:

team: (str) The name of the team
admin_username: (str) optional username of the admin user of the team, defaults to the current user.

Returns: A Team object

`method` `Api.create_user`

create_user(email, admin=False)

Create a new user.

Args:

email: (str) The email address of the user
admin: (bool) Whether this user should be a global instance admin

Returns: A User object

`method` `Api.flush`

flush()

Flush the local cache.

The api object keeps a local cache of runs, so if the state of the run may change while executing your script you must clear the local cache with api.flush() to get the latest values associated with the run.

`method` `Api.from_path`

from_path(path)

Return a run, sweep, project or report from a path.

Examples:

    project = api.from_path("my_project")
    team_project = api.from_path("my_team/my_project")
    run = api.from_path("my_team/my_project/runs/id")
    sweep = api.from_path("my_team/my_project/sweeps/id")
    report = api.from_path("my_team/my_project/reports/My-Report-Vm11dsdf")
   ``` 



**Args:**

- `path`:  (str) The path to the project, run, sweep or report 



**Returns:**
A `Project`, `Run`, `Sweep`, or `BetaReport` instance. 



**Raises:**
wandb.Error if path is invalid or the object doesn't exist 

---

### <kbd>method</kbd> `Api.job`

```python
job(name: Optional[str], path: Optional[str] = None) → public.Job

Return a Job from the given parameters.

Args:

name: (str) The job name.
path: (str, optional) If given, the root path in which to download the job artifact.

Returns: A Job object.

`method` `Api.list_jobs`

list_jobs(entity: str, project: str) → List[Dict[str, Any]]

Return a list of jobs, if any, for the given entity and project.

Args:

entity: (str) The entity for the listed job(s).
project: (str) The project for the listed job(s).

Returns: A list of matching jobs.

`method` `Api.project`

project(name: str, entity: Optional[str] = None) → public.Project

Return the Project with the given name (and entity, if given).

Args:

name: (str) The project name.
entity: (str) Name of the entity requested. If None, will fall back to the default entity passed to Api. If no default entity, will raise a ValueError.

Returns: A Project object.

`method` `Api.projects`

projects(entity: Optional[str] = None, per_page: int = 200) → public.Projects

Get projects for a given entity.

Args:

entity: (str) Name of the entity requested. If None, will fall back to the default entity passed to Api. If no default entity, will raise a ValueError.
per_page: (int) Sets the page size for query pagination. Usually there is no reason to change this.

Returns: A Projects object which is an iterable collection of Project objects.

`method` `Api.queued_run`

queued_run(
    entity,
    project,
    queue_name,
    run_queue_item_id,
    project_queue=None,
    priority=None
)

Return a single queued run based on the path.

Parses paths of the form entity/project/queue_id/run_queue_item_id.

`method` `Api.registries`

registries(
    organization: Optional[str] = None,
    filter: Optional[Dict[str, Any]] = None
) → Registries

Returns a Registry iterator.

Use the iterator to search and filter registries, collections, or artifact versions across your organization’s registry.

Examples: Find all registries with the names that contain “model” ```python import wandb

 api = wandb.Api()  # specify an org if your entity belongs to multiple orgs
 api.registries(filter={"name": {"$regex": "model"}})
```

Find all collections in the registries with the name “my_collection” and the tag “my_tag” python api.registries().collections(filter={"name": "my_collection", "tag": "my_tag"})

Find all artifact versions in the registries with a collection name that contains “my_collection” and a version that has the alias “best” python api.registries().collections( filter={"name": {"$regex": "my_collection"}} ).versions(filter={"alias": "best"})

Find all artifact versions in the registries that contain “model” and have the tag “prod” or alias “best” python api.registries(filter={"name": {"$regex": "model"}}).versions( filter={"$or": [{"tag": "prod"}, {"alias": "best"}]} )

Args:

organization: (str, optional) The organization of the registry to fetch. If not specified, use the organization specified in the user’s settings.
filter: (dict, optional) MongoDB-style filter to apply to each object in the registry iterator. Fields available to filter for collections are name, description, created_at, updated_at. Fields available to filter for collections are name, tag, description, created_at, updated_at Fields available to filter for versions are tag, alias, created_at, updated_at, metadata

Returns: A registry iterator.

`method` `Api.reports`

reports(
    path: str = '',
    name: Optional[str] = None,
    per_page: int = 50
) → public.Reports

Get reports for a given project path.

WARNING: This api is in beta and will likely change in a future release

Args:

path: (str) path to project the report resides in, should be in the form: “entity/project”
name: (str, optional) optional name of the report requested.
per_page: (int) Sets the page size for query pagination. Usually there is no reason to change this.

Returns: A Reports object which is an iterable collection of BetaReport objects.

`method` `Api.run`

run(path='')

Return a single run by parsing path in the form entity/project/run_id.

Args:

path: (str) path to run in the form entity/project/run_id. If api.entity is set, this can be in the form project/run_id and if api.project is set this can just be the run_id.

Returns: A Run object.

`method` `Api.run_queue`

run_queue(entity, name)

Return the named RunQueue for entity.

To create a new RunQueue, use wandb.Api().create_run_queue(...).

`method` `Api.runs`

runs(
    path: Optional[str] = None,
    filters: Optional[Dict[str, Any]] = None,
    order: str = '+created_at',
    per_page: int = 50,
    include_sweeps: bool = True
)

Return a set of runs from a project that match the filters provided.

Fields you can filter by include:

createdAt: The timestamp when the run was created. (in ISO 8601 format, e.g. “2023-01-01T12:00:00Z”)
displayName: The human-readable display name of the run. (e.g. “eager-fox-1”)
duration: The total runtime of the run in seconds.
group: The group name used to organize related runs together.
host: The hostname where the run was executed.
jobType: The type of job or purpose of the run.
name: The unique identifier of the run. (e.g. “a1b2cdef”)
state: The current state of the run.
tags: The tags associated with the run.
username: The username of the user who initiated the run

Additionally, you can filter by items in the run config or summary metrics. Such as config.experiment_name, summary_metrics.loss, etc.

For more complex filtering, you can use MongoDB query operators. For details, see: https://docs.mongodb.com/manual/reference/operator/query The following operations are supported:

$and
$or
$nor
$eq
$ne
$gt
$gte
$lt
$lte
$in
$nin
$exists
$regex

Examples: Find runs in my_project where config.experiment_name has been set to “foo” api.runs( path="my_entity/my_project", filters={"config.experiment_name": "foo"}, )

Find runs in my_project where config.experiment_name has been set to “foo” or “bar” api.runs( path="my_entity/my_project", filters={ "$or": [ {"config.experiment_name": "foo"}, {"config.experiment_name": "bar"}, ] }, )

Find runs in my_project where config.experiment_name matches a regex (anchors are not supported) api.runs( path="my_entity/my_project", filters={"config.experiment_name": {"$regex": "b.*"}}, )

Find runs in my_project where the run name matches a regex (anchors are not supported) api.runs( path="my_entity/my_project", filters={"display_name": {"$regex": "^foo.*"}}, )

Find runs in my_project where config.experiment contains a nested field “category” with value “testing” api.runs( path="my_entity/my_project", filters={"config.experiment.category": "testing"}, )

Find runs in my_project with a loss value of 0.5 nested in a dictionary under model1 in the summary metrics api.runs( path="my_entity/my_project", filters={"summary_metrics.model1.loss": 0.5}, )

Find runs in my_project sorted by ascending loss api.runs(path="my_entity/my_project", order="+summary_metrics.loss")

Args:

path: (str) path to project, should be in the form: “entity/project”
filters: (dict) queries for specific runs using the MongoDB query language. You can filter by run properties such as config.key, summary_metrics.key, state, entity, createdAt, etc.
For example: {"config.experiment_name": "foo"} would find runs with a config entry of experiment name set to “foo”
order: (str) Order can be created_at, heartbeat_at, config.*.value, or summary_metrics.*. If you prepend order with a + order is ascending. If you prepend order with a - order is descending (default). The default order is run.created_at from oldest to newest.
per_page: (int) Sets the page size for query pagination.
include_sweeps: (bool) Whether to include the sweep runs in the results.

Returns: A Runs object, which is an iterable collection of Run objects.

`method` `Api.sweep`

sweep(path='')

Return a sweep by parsing path in the form entity/project/sweep_id.

Args:

path: (str, optional) path to sweep in the form entity/project/sweep_id. If api.entity is set, this can be in the form project/sweep_id and if api.project is set this can just be the sweep_id.

Returns: A Sweep object.

`method` `Api.sync_tensorboard`

sync_tensorboard(root_dir, run_id=None, project=None, entity=None)

Sync a local directory containing tfevent files to wandb.

`method` `Api.team`

team(team: str) → public.Team

Return the matching Team with the given name.

Args:

team: (str) The name of the team.

Returns: A Team object.

`method` `Api.upsert_run_queue`

upsert_run_queue(
    name: str,
    resource_config: dict,
    resource_type: 'public.RunQueueResourceType',
    entity: Optional[str] = None,
    template_variables: Optional[dict] = None,
    external_links: Optional[dict] = None,
    prioritization_mode: Optional[ForwardRef('public.RunQueuePrioritizationMode')] = None
)

Upsert a run queue (launch).

Args:

name: (str) Name of the queue to create
entity: (str) Optional name of the entity to create the queue. If None, will use the configured or default entity.
resource_config: (dict) Optional default resource configuration to be used for the queue. Use handlebars (eg. {{var}}) to specify template variables.
resource_type: (str) Type of resource to be used for the queue. One of “local-container”, “local-process”, “kubernetes”, “sagemaker”, or “gcp-vertex”.
template_variables: (dict) A dictionary of template variable schemas to be used with the config. Expected format of: `{
"var-name": {
"schema": {
"type": (“string”, “number”, or “integer”),
"default": (optional value),
"minimum": (optional minimum),
"maximum": (optional maximum),
"enum": […"(options)"] } } }`
external_links: (dict) Optional dictionary of external links to be used with the queue. Expected format of: `{
"name": “url” }`
prioritization_mode: (str) Optional version of prioritization to use. Either “V0” or None

Returns: The upserted RunQueue.

Raises: ValueError if any of the parameters are invalid wandb.Error on wandb API errors

`method` `Api.user`

user(username_or_email: str) → Optional[ForwardRef('public.User')]

Return a user from a username or email address.

Note: This function only works for Local Admins, if you are trying to get your own user object, please use api.viewer.

Args:

username_or_email: (str) The username or email address of the user

Returns: A User object or None if a user couldn’t be found

`method` `Api.users`

users(username_or_email: str) → List[ForwardRef('public.User')]

Return all users from a partial username or email address query.

Note: This function only works for Local Admins, if you are trying to get your own user object, please use api.viewer.

Args:

username_or_email: (str) The prefix or suffix of the user you want to find

Returns: An array of User objects

2 - artifacts

View the source code

`module` `wandb.apis.public`

W&B Public API for Artifact Management.

This module provides classes for interacting with W&B artifacts and their collections. Classes include:

ArtifactTypes: A paginated collection of artifact types in a project

List and query artifact types
Access type metadata
Create new types

ArtifactCollection: A collection of related artifacts

Manage artifact collections
Update metadata and descriptions
Work with tags and aliases
Change collection types

Artifacts: A paginated collection of artifact versions

Filter and query artifacts
Access artifact metadata
Download artifacts

ArtifactFiles: A paginated collection of files within an artifact

List and query artifact files
Access file metadata
Download individual files

`function` `server_supports_artifact_collections_gql_edges`

server_supports_artifact_collections_gql_edges(
    client: 'RetryingClient',
    warn: bool = False
) → bool

Check if W&B server supports GraphQL edges for artifact collections.

`function` `artifact_collection_edge_name`

artifact_collection_edge_name(server_supports_artifact_collections: bool) → str

Return the GraphQL edge name for artifact collections or sequences.

`function` `artifact_collection_plural_edge_name`

artifact_collection_plural_edge_name(
    server_supports_artifact_collections: bool
) → str

Return the GraphQL edge name for artifact collections or sequences.

`class` `ArtifactTypes`

An iterable collection of artifact types associated with a project.

Args:

client: The client instance to use for querying W&B.
entity: The entity (user or team) that owns the project.
project: The name of the project to query for artifact types.
per_page: The number of artifact types to fetch per page. Default is 50.

`method` `ArtifactTypes.init`

__init__(
    client: wandb_gql.client.Client,
    entity: str,
    project: str,
    per_page: int = 50
)

`property` ArtifactTypes.cursor

Returns the cursor position for pagination of file results.

`property` ArtifactTypes.length

Returns None.

`property` ArtifactTypes.more

Returns True if there are more artifacts to fetch. Returns False if there are no more files to fetch.

`method` `ArtifactTypes.convert_objects`

convert_objects()

Converts GraphQL edges to ArtifactType objects.

`method` `ArtifactTypes.update_variables`

update_variables()

Updates the variables dictionary with the cursor.

`class` `ArtifactType`

An artifact object that satisfies query based on the specified type.

Args:

client: The client instance to use for querying W&B.
entity: The entity (user or team) that owns the project.
project: The name of the project to query for artifact types.
type_name: The name of the artifact type.
attrs: Optional mapping of attributes to initialize the artifact type. If not provided, the object will load its attributes from W&B upon initialization.

`method` `ArtifactType.init`

__init__(
    client: wandb_gql.client.Client,
    entity: str,
    project: str,
    type_name: str,
    attrs: Optional[Mapping[str, Any]] = None
)

`property` ArtifactType.id

The unique identifier of the artifact type.

`property` ArtifactType.name

The name of the artifact type.

`method` `ArtifactType.collection`

collection(name)

Get a specific artifact collection by name.

Args:

name (str): The name of the artifact collection to retrieve.

`method` `ArtifactType.collections`

collections(per_page=50)

Get all artifact collections associated with this artifact type.

Args:

per_page (int): The number of artifact collections to fetch per page. Default is 50.

`method` `ArtifactType.load`

load()

Load the artifact type metadata.

`class` `ArtifactCollections`

An iterable collection of artifact collections associated with a project and artifact type.

Args:

client: The client instance to use for querying W&B.
entity: The entity (user or team) that owns the project.
project: The name of the project to query for artifact collections.
type_name: The name of the artifact type for which to fetch collections.
per_page: The number of artifact collections to fetch per page. Default is 50.

`method` `ArtifactCollections.init`

__init__(
    client: wandb_gql.client.Client,
    entity: str,
    project: str,
    type_name: str,
    per_page: int = 50
)

`property` ArtifactCollections.cursor

Returns the cursor position for pagination of file results.

`property` ArtifactCollections.length

Returns the number of artifact collections.

`property` ArtifactCollections.more

Returns True if there are more artifact collections to fetch.

Returns False if there are no more files to fetch.

`method` `ArtifactCollections.convert_objects`

convert_objects()

Converts GraphQL edges to ArtifactCollection objects.

`method` `ArtifactCollections.update_variables`

update_variables()

Updates the variables dictionary with the cursor.

`class` `ArtifactCollection`

An artifact collection that represents a group of related artifacts.

Args:

client: The client instance to use for querying W&B.
entity: The entity (user or team) that owns the project.
project: The name of the project to query for artifact collections.
name: The name of the artifact collection.
type: The type of the artifact collection (e.g., “dataset”, “model”).
organization: Optional organization name if applicable.
attrs: Optional mapping of attributes to initialize the artifact collection. If not provided, the object will load its attributes from W&B upon initialization.

`method` `ArtifactCollection.init`

__init__(
    client: wandb_gql.client.Client,
    entity: str,
    project: str,
    name: str,
    type: str,
    organization: Optional[str] = None,
    attrs: Optional[Mapping[str, Any]] = None
)

`property` ArtifactCollection.aliases

The aliases associated with the artifact collection.

`property` ArtifactCollection.created_at

The creation timestamp of the artifact collection.

`property` ArtifactCollection.description

A description of the artifact collection.

`property` ArtifactCollection.id

The unique identifier of the artifact collection.

`property` ArtifactCollection.name

The name of the artifact collection.

`property` ArtifactCollection.tags

The tags associated with the artifact collection.

`property` ArtifactCollection.type

Returns the type of the artifact collection.

`method` `ArtifactCollection.artifacts`

artifacts(per_page=50)

Get all artifact versions associated with this artifact collection.

`method` `ArtifactCollection.change_type`

change_type(new_type: str) → None

Deprecated, change type directly with save instead.

`method` `ArtifactCollection.delete`

delete()

Delete the entire artifact collection.

`method` `ArtifactCollection.is_sequence`

is_sequence() → bool

Return whether the artifact collection is a sequence.

`method` `ArtifactCollection.load`

load()

Load the artifact collection metadata.

`method` `ArtifactCollection.save`

save() → None

Persist any changes made to the artifact collection.

`class` `Artifacts`

An iterable collection of artifact versions associated with a project.

Optionally pass in filters to narrow down the results based on specific criteria.

Args:

client: The client instance to use for querying W&B.
entity: The entity (user or team) that owns the project.
project: The name of the project to query for artifacts.
collection_name: The name of the artifact collection to query.
type: The type of the artifacts to query. Common examples include “dataset” or “model”.
filters: Optional mapping of filters to apply to the query.
order: Optional string to specify the order of the results.
per_page: The number of artifact versions to fetch per page. Default is 50.
tags: Optional string or list of strings to filter artifacts by tags.

`method` `Artifacts.init`

__init__(
    client: wandb_gql.client.Client,
    entity: str,
    project: str,
    collection_name: str,
    type: str,
    filters: Optional[Mapping[str, Any]] = None,
    order: Optional[str] = None,
    per_page: int = 50,
    tags: Optional[str, List[str]] = None
)

`property` Artifacts.cursor

Returns the cursor position for pagination of file results.

`property` Artifacts.length

Returns the number of artifact versions.

`property` Artifacts.more

Returns True if there are more artifact versions to fetch.

`method` `Artifacts.convert_objects`

convert_objects()

Converts GraphQL edges to Artifact objects.

`class` `RunArtifacts`

An iterable collection of artifacts associated with a run.

Args:

client: The client instance to use for querying W&B.
run: The run object to query for artifacts.
mode: The mode of artifacts to fetch, either “logged” (output artifacts) or “used” (input artifacts). Default is “logged”.
per_page: The number of artifacts to fetch per page. Default is 50.

`method` `RunArtifacts.init`

__init__(
    client: wandb_gql.client.Client,
    run: 'Run',
    mode='logged',
    per_page: int = 50
)

`property` RunArtifacts.cursor

Returns the cursor position for pagination of file results.

`property` RunArtifacts.length

Returns the number of artifacts associated with the run.

`property` RunArtifacts.more

Returns True if there are more artifacts to fetch.

`method` `RunArtifacts.convert_objects`

convert_objects()

Converts GraphQL edges to Artifact objects.

`class` `ArtifactFiles`

An iterable collection of files associated with an artifact version.

Args:

client: The client instance to use for querying W&B.
artifact: The artifact object to query for files.
names: Optional sequence of file names to filter the results by. If None, all files will be returned.
per_page: The number of files to fetch per page. Default is 50.

`method` `ArtifactFiles.init`

__init__(
    client: wandb_gql.client.Client,
    artifact: 'wandb.Artifact',
    names: Optional[Sequence[str]] = None,
    per_page: int = 50
)

`property` ArtifactFiles.cursor

Returns the cursor position for pagination of file results.

`property` ArtifactFiles.length

Returns the number of files in the artifact.

`property` ArtifactFiles.more

Returns True if there are more files to fetch. Returns

`property` ArtifactFiles.path

Returns the path of the artifact.

The path is a list containingthe entity, project name, and artifact name.

`method` `ArtifactFiles.convert_objects`

convert_objects()

Converts GraphQL edges to File objects.

`method` `ArtifactFiles.update_variables`

update_variables()

Updates the variables dictionary with the cursor and limit.

3 - files

View the source code

`module` `wandb.apis.public`

W&B Public API for File Management.

This module provides classes for interacting with files stored in W&B. Classes include:

Files: A paginated collection of files associated with a run

Iterate through files with automatic pagination
Filter files by name
Access file metadata and properties
Download multiple files

File: A single file stored in W&B

Access file metadata (size, mimetype, URLs)
Download files to local storage
Delete files from W&B
Work with S3 URIs for direct access

Example:

from wandb.apis.public import Api

# Initialize API
api = Api()

# Get files from a specific run
run = api.run("entity/project/run_id")
files = run.files()

# Work with files
for file in files:
    print(f"File: {file.name}")
    print(f"Size: {file.size} bytes")
    print(f"Type: {file.mimetype}")

    # Download file
    if file.size < 1000000:  # Less than 1MB
        file.download(root="./downloads")

    # Get S3 URI for large files
    if file.size >= 1000000:
        print(f"S3 URI: {file.path_uri}")

Note:

This module is part of the W&B Public API and provides methods to access, download, and manage files stored in W&B. Files are typically associated with specific runs and can include model weights, datasets, visualizations, and other artifacts.

`class` `Files`

An iterable collection of File objects.

Access and manage files uploaded to W&B during a run. Handles pagination automatically when iterating through large collections of files.

Args:

client: The run object that contains the files
run: The run object that contains the files
names (list, optional): A list of file names to filter the files
per_page (int, optional): The number of files to fetch per page
upload (bool, optional): If True, fetch the upload URL for each file

Example:

from wandb.apis.public.files import Files
from wandb.apis.public.api import Api

# Initialize the API client
api = Api()

# Example run object
run = api.run("entity/project/run-id")

# Create a Files object to iterate over files in the run
files = Files(api.client, run)

# Iterate over files
for file in files:
   print(file.name)
   print(file.url)
   print(file.size)

   # Download the file
   file.download(root="download_directory", replace=True)

`method` `Files.init`

__init__(client, run, names=None, per_page=50, upload=False)

`property` Files.cursor

Returns the cursor position for pagination of file results.

`property` Files.length

The number of files saved to the specified run.

`property` Files.more

Returns True if there are more files to fetch. Returns False if there are no more files to fetch.

`method` `Files.convert_objects`

convert_objects()

Converts GraphQL edges to File objects.

`method` `Files.update_variables`

update_variables()

Updates the GraphQL query variables for pagination.

`class` `File`

File saved to W&B.

Represents a single file stored in W&B. Includes access to file metadata. Files are associated with a specific run and can include text files, model weights, datasets, visualizations, and other artifacts. You can download the file, delete the file, and access file properties.

Specify one or more attributes in a dictionary to fine a specific file logged to a specific run. You can search using the following keys:

id (str): The ID of the run that contains the file
name (str): Name of the file
url (str): path to file
direct_url (str): path to file in the bucket
sizeBytes (int): size of file in bytes
md5 (str): md5 of file
mimetype (str): mimetype of file
updated_at (str): timestamp of last update
path_uri (str): path to file in the bucket, currently only available for files stored in S3

Args:

client: The run object that contains the file
attrs (dict): A dictionary of attributes that define the file
run: The run object that contains the file

Example:

from wandb.apis.public.files import File
from wandb.apis.public.api import Api

# Initialize the API client
api = Api()

# Example attributes dictionary
file_attrs = {
   "id": "file-id",
   "name": "example_file.txt",
   "url": "https://example.com/file",
   "direct_url": "https://example.com/direct_file",
   "sizeBytes": 1024,
   "mimetype": "text/plain",
   "updated_at": "2025-03-25T21:43:51Z",
   "md5": "d41d8cd98f00b204e9800998ecf8427e",
}

# Example run object
run = api.run("entity/project/run-id")

# Create a File object
file = File(api.client, file_attrs, run)

# Access some of the attributes
print("File ID:", file.id)
print("File Name:", file.name)
print("File URL:", file.url)
print("File MIME Type:", file.mimetype)
print("File Updated At:", file.updated_at)

# Access File properties
print("File Size:", file.size)
print("File Path URI:", file.path_uri)

# Download the file
file.download(root="download_directory", replace=True)

# Delete the file
file.delete()

`method` `File.init`

__init__(client, attrs, run=None)

`property` File.path_uri

Returns the URI path to the file in the storage bucket.

`property` File.size

Returns the size of the file in bytes.

`method` `File.delete`

delete()

Deletes the file from the W&B server.

`method` `File.download`

download(
    root: str = '.',
    replace: bool = False,
    exist_ok: bool = False,
    api: Optional[wandb.apis.public.api.Api] = None
) → TextIOWrapper

Downloads a file previously saved by a run from the wandb server.

Args:

root: Local directory to save the file. Defaults to “.”.
replace: If True, download will overwrite a local file if it exists. Defaults to False.
exist_ok: If True, will not raise ValueError if file already exists and will not re-download unless replace=True. Defaults to False.
api: If specified, the Api instance used to download the file.

Raises: ValueError if file already exists, replace=False and exist_ok=False.

4 - history

View the source code

`module` `wandb.apis.public`

W&B Public API for Run History.

This module provides classes for efficiently scanning and sampling run history data. Classes include:

HistoryScan: Iterator for scanning complete run history

Paginated access to all metrics
Configure step ranges and page sizes
Raw access to all logged data

SampledHistoryScan: Iterator for sampling run history data

Efficient access to downsampled metrics
Filter by specific keys
Control sample size and step ranges

Note:

This module is part of the W&B Public API and provides methods to access run history data. It handles pagination automatically and offers both complete and sampled access to metrics logged during training runs.

`class` `HistoryScan`

Iterator for scanning complete run history.

Args:

client: (wandb.apis.internal.Api) The client instance to use
run: (wandb.sdk.internal.Run) The run object to scan history for
min_step: (int) The minimum step to start scanning from
max_step: (int) The maximum step to scan up to
page_size: (int) Number of samples per page (default is 1000)

`method` `HistoryScan.init`

__init__(client, run, min_step, max_step, page_size=1000)

`class` `SampledHistoryScan`

Iterator for sampling run history data.

Args:

client: (wandb.apis.internal.Api) The client instance to use
run: (wandb.sdk.internal.Run) The run object to sample history from
keys: (list) List of keys to sample from the history
min_step: (int) The minimum step to start sampling from
max_step: (int) The maximum step to sample up to
page_size: (int) Number of samples per page (default is 1000)

`method` `SampledHistoryScan.init`

__init__(client, run, keys, min_step, max_step, page_size=1000)

5 - jobs

View the source code

`module` `wandb.apis.public`

W&B Public API for Job Management and Queuing.

This module provides classes for managing W&B jobs, queued runs, and run queues. Classes include:

Job: Manage W&B job definitions and execution

Load and configure jobs from artifacts
Set entrypoints and runtime configurations
Execute jobs with different resource types
Handle notebook and container-based jobs

QueuedRun: Track and manage individual queued runs

Monitor run state and execution
Wait for run completion
Access run results and artifacts
Delete queued runs

RunQueue: Manage job queues and execution resources

Create and configure run queues
Set resource types and configurations
Monitor queue items and status
Control queue access and priorities

`class` `Job`

`method` `Job.init`

__init__(api: 'Api', name, path: Optional[str] = None) → None

`property` Job.name

The name of the job.

`method` `Job.call`

call(
    config,
    project=None,
    entity=None,
    queue=None,
    resource='local-container',
    resource_args=None,
    template_variables=None,
    project_queue=None,
    priority=None
)

Call the job with the given configuration.

Args:

config (dict): The configuration to pass to the job. This should be a dictionary containing key-value pairs that match the input types defined in the job.
project (str, optional): The project to log the run to. Defaults to the job’s project.
entity (str, optional): The entity to log the run under. Defaults to the job’s entity.
queue (str, optional): The name of the queue to enqueue the job to. Defaults to None.
resource (str, optional): The resource type to use for execution. Defaults to “local-container”.
resource_args (dict, optional): Additional arguments for the resource type. Defaults to None.
template_variables (dict, optional): Template variables to use for the job. Defaults to None.
project_queue (str, optional): The project that manages the queue. Defaults to None.
priority (int, optional): The priority of the queued run. Defaults to None.

`method` `Job.set_entrypoint`

set_entrypoint(entrypoint: List[str])

Set the entrypoint for the job.

`class` `QueuedRun`

A single queued run associated with an entity and project.

Args:

entity: The entity associated with the queued run.
project (str): The project where runs executed by the queue are logged to.
queue_name (str): The name of the queue.
run_queue_item_id (int): The id of the run queue item.
project_queue (str): The project that manages the queue.
priority (str): The priority of the queued run.

Call run = queued_run.wait_until_running() or run = queued_run.wait_until_finished() to access the run.

`method` `QueuedRun.init`

__init__(
    client,
    entity,
    project,
    queue_name,
    run_queue_item_id,
    project_queue='model-registry',
    priority=None
)

`property` QueuedRun.entity

The entity associated with the queued run.

`property` QueuedRun.id

The id of the queued run.

`property` QueuedRun.project

The project associated with the queued run.

`property` QueuedRun.queue_name

The name of the queue.

`property` QueuedRun.state

The state of the queued run.

`method` `QueuedRun.delete`

delete(delete_artifacts=False)

Delete the given queued run from the wandb backend.

`method` `QueuedRun.wait_until_finished`

wait_until_finished()

Wait for the queued run to complete and return the finished run.

`method` `QueuedRun.wait_until_running`

wait_until_running()

Wait until the queued run is running and return the run.

`class` `RunQueue`

Class that represents a run queue in W&B.

Args:

client: W&B API client instance.
name: Name of the run queue
entity: The entity (user or team) that owns this queue
prioritization_mode: Queue priority mode Can be “DISABLED” or “V0”. Defaults to None.
_access: Access level for the queue Can be “project” or “user”. Defaults to None.
_default_resource_config_id: ID of default resource config
_default_resource_config: Default resource configuration

`method` `RunQueue.init`

__init__(
    client: 'RetryingClient',
    name: str,
    entity: str,
    prioritization_mode: Optional[Literal['DISABLED', 'V0']] = None,
    _access: Optional[Literal['project', 'user']] = None,
    _default_resource_config_id: Optional[int] = None,
    _default_resource_config: Optional[dict] = None
) → None

`property` RunQueue.access

The access level of the queue.

`property` RunQueue.default_resource_config

The default configuration for resources.

`property` RunQueue.entity

The entity that owns the queue.

`property` RunQueue.external_links

External resource links for the queue.

`property` RunQueue.id

The id of the queue.

`property` RunQueue.items

Up to the first 100 queued runs. Modifying this list will not modify the queue or any enqueued items!

`property` RunQueue.name

The name of the queue.

`property` RunQueue.prioritization_mode

The prioritization mode of the queue.

Can be set to “DISABLED” or “V0”.

`property` RunQueue.template_variables

Variables for resource templates.

`property` RunQueue.type

The resource type for execution.

`classmethod` `RunQueue.create`

create(
    name: str,
    resource: 'RunQueueResourceType',
    entity: Optional[str] = None,
    prioritization_mode: Optional[ForwardRef('RunQueuePrioritizationMode')] = None,
    config: Optional[dict] = None,
    template_variables: Optional[dict] = None
) → RunQueue

Create a RunQueue.

Args:

name: The name of the run queue to create.
resource: The resource type for execution.
entity: The entity (user or team) that will own the queue. Defaults to the default entity of the API client.
prioritization_mode: The prioritization mode for the queue. Can be “DISABLED” or “V0”. Defaults to None.
config: Optional dictionary for the default resource configuration. Defaults to None.
template_variables: Optional dictionary for template variables used in the resource configuration.

`method` `RunQueue.delete`

delete()

Delete the run queue from the wandb backend.

6 - projects

View the source code

`module` `wandb.apis.public`

W&B Public API for Projects.

This module provides classes for interacting with W&B projects and their associated data. Classes include:

Projects: A paginated collection of projects associated with an entity

Iterate through all projects
Access project metadata
Query project information

Project: A single project that serves as a namespace for runs

Access project properties
Work with artifacts and their types
Manage sweeps
Generate HTML representations for Jupyter

Example:

from wandb.apis.public import Api

# Initialize API
api = Api()

# Get all projects for an entity
projects = api.projects("entity")

# Access project data
for project in projects:
    print(f"Project: {project.name}")
    print(f"URL: {project.url}")

    # Get artifact types
    for artifact_type in project.artifacts_types():
        print(f"Artifact Type: {artifact_type.name}")

    # Get sweeps
    for sweep in project.sweeps():
        print(f"Sweep ID: {sweep.id}")
        print(f"State: {sweep.state}")

Note:

This module is part of the W&B Public API and provides methods to access and manage projects. For creating new projects, use wandb.init() with a new project name.

`class` `Projects`

An iterable collection of Project objects.

An iterable interface to access projects created and saved by the entity.

Args:

client (wandb.apis.internal.Api): The API client instance to use.
entity (str): The entity name (username or team) to fetch projects for.
per_page (int): Number of projects to fetch per request (default is 50).

Example:

from wandb.apis.public.api import Api

# Initialize the API client
api = Api()

# Find projects that belong to this entity
projects = api.projects(entity="entity")

# Iterate over files
for project in projects:
   print(f"Project: {project.name}")
   print(f"- URL: {project.url}")
   print(f"- Created at: {project.created_at}")
   print(f"- Is benchmark: {project.is_benchmark}")

`method` `Projects.init`

__init__(client, entity, per_page=50)

`property` Projects.cursor

Returns the cursor position for pagination of project results.

`property` Projects.length

Returns the total number of projects.

Note: This property is not available for projects.

`property` Projects.more

Returns True if there are more projects to fetch. Returns False if there are no more projects to fetch.

`method` `Projects.convert_objects`

convert_objects()

Converts GraphQL edges to File objects.

`class` `Project`

A project is a namespace for runs.

Args:

client: W&B API client instance.
name (str): The name of the project.
entity (str): The entity name that owns the project.

`method` `Project.init`

__init__(client, entity, project, attrs)

`property` Project.path

Returns the path of the project. The path is a list containing the entity and project name.

`property` Project.url

Returns the URL of the project.

`method` `Project.artifacts_types`

artifacts_types(per_page=50)

Returns all artifact types associated with this project.

`method` `Project.sweeps`

sweeps()

Fetches all sweeps associated with the project.

`method` `Project.to_html`

to_html(height=420, hidden=False)

Generate HTML containing an iframe displaying this project.

7 - query_generator

View the source code

`module` `wandb.apis.public`

`method` `QueryGenerator.filter_to_mongo`

filter_to_mongo(filter)

Returns dictionary with filter format converted to MongoDB filter.

`classmethod` `QueryGenerator.format_order_key`

format_order_key(key: str)

Format a key for sorting.

`method` `QueryGenerator.key_to_server_path`

key_to_server_path(key)

Convert a key dictionary to the corresponding server path string.

`method` `QueryGenerator.keys_to_order`

keys_to_order(keys)

Convert a list of key dictionaries to an order string.

`method` `QueryGenerator.mongo_to_filter`

mongo_to_filter(filter)

Returns dictionary with MongoDB filter converted to filter format.

`method` `QueryGenerator.order_to_keys`

order_to_keys(order)

Convert an order string to a list of key dictionaries.

`method` `QueryGenerator.server_path_to_key`

server_path_to_key(path)

Convert a server path string to the corresponding key dictionary.

8 - reports

View the source code

`module` `wandb.apis.public`

Public API: reports.

`class` `Reports`

Reports is an iterable collection of BetaReport objects.

Args:

client (wandb.apis.internal.Api): The API client instance to use.
project (wandb.sdk.internal.Project): The project to fetch reports from.
name (str, optional): The name of the report to filter by. If None, fetches all reports.
entity (str, optional): The entity name for the project. Defaults to the project entity.
per_page (int): Number of reports to fetch per page (default is 50).

`method` `Reports.init`

__init__(client, project, name=None, entity=None, per_page=50)

`property` Reports.cursor

Returns the cursor position for pagination of file results.

`property` Reports.length

The number of reports in the project.

`property` Reports.more

Returns True if there are more files to fetch. Returns False if there are no more files to fetch.

`method` `Reports.convert_objects`

convert_objects()

Converts GraphQL edges to File objects.

`method` `Reports.update_variables`

update_variables()

Updates the GraphQL query variables for pagination.

`class` `BetaReport`

BetaReport is a class associated with reports created in wandb.

WARNING: this API will likely change in a future release

Attributes:

name (string): report name
description (string): report description
user (User): the user that created the report
spec (dict): the spec off the report
updated_at (string): timestamp of last update

`method` `BetaReport.init`

__init__(client, attrs, entity=None, project=None)

`property` BetaReport.sections

Get the panel sections (groups) from the report.

`property` BetaReport.updated_at

Timestamp of last update

`property` BetaReport.url

URL of the report.

Contains the entity, project, display name, and id.

`method` `BetaReport.runs`

runs(section, per_page=50, only_selected=True)

Get runs associated with a section of the report.

`method` `BetaReport.to_html`

to_html(height=1024, hidden=False)

Generate HTML containing an iframe displaying this report.

9 - runs

View the source code

`module` `wandb.apis.public`

W&B Public API for ML Runs.

This module provides classes for interacting with W&B runs and their associated data. Classes include:

Runs: A paginated collection of runs associated with a project

Filter and query runs
Access run histories and metrics
Export data in various formats (pandas, polars)

Run: A single machine learning training run

Access run metadata, configs, and metrics
Upload and download files
Work with artifacts
Query run history
Update run information

Example:

from wandb.apis.public import Api

# Initialize API
api = Api()

# Get runs matching filters
runs = api.runs(
    path="entity/project", filters={"state": "finished", "config.batch_size": 32}
)

# Access run data
for run in runs:
    print(f"Run: {run.name}")
    print(f"Config: {run.config}")
    print(f"Metrics: {run.summary}")

    # Get history with pandas
    history_df = run.history(keys=["loss", "accuracy"], pandas=True)

    # Work with artifacts
    for artifact in run.logged_artifacts():
        print(f"Artifact: {artifact.name}")

Note:

This module is part of the W&B Public API and provides read/write access to run data. For logging new runs, use the wandb.init() function from the main wandb package.

`class` `Runs`

An iterable collection of runs associated with a project and optional filter.

This is generally used indirectly using the Api.runs namespace.

Args:

client: (wandb.apis.public.RetryingClient) The API client to use for requests.
entity: (str) The entity (username or team) that owns the project.
project: (str) The name of the project to fetch runs from.
filters: (Optional[Dict[str, Any]]) A dictionary of filters to apply to the runs query.
order: (Optional[str]) The order of the runs, can be “asc” or “desc” Defaults to “desc”.
per_page: (int) The number of runs to fetch per request (default is 50).
include_sweeps: (bool) Whether to include sweep information in the runs. Defaults to True.

Examples:

from wandb.apis.public.runs import Runs
from wandb.apis.public import Api

# Initialize the API client
api = Api()

# Get all runs from a project that satisfy the filters
filters = {"state": "finished", "config.optimizer": "adam"}

runs = Runs(
   client=api.client,
   entity="entity",
   project="project_name",
   filters=filters,
)

# Iterate over runs and print details
for run in runs:
   print(f"Run name: {run.name}")
   print(f"Run ID: {run.id}")
   print(f"Run URL: {run.url}")
   print(f"Run state: {run.state}")
   print(f"Run config: {run.config}")
   print(f"Run summary: {run.summary}")
   print(f"Run history (samples=5): {run.history(samples=5)}")
   print("----------")

# Get histories for all runs with specific metrics
histories_df = runs.histories(
   samples=100,  # Number of samples per run
   keys=["loss", "accuracy"],  # Metrics to fetch
   x_axis="_step",  # X-axis metric
   format="pandas",  # Return as pandas DataFrame
)

`method` `Runs.init`

__init__(
    client: 'RetryingClient',
    entity: str,
    project: str,
    filters: Optional[Dict[str, Any]] = None,
    order: Optional[str] = None,
    per_page: int = 50,
    include_sweeps: bool = True
)

`property` Runs.cursor

Returns the cursor position for pagination of runs results.

`property` Runs.length

Returns the total number of runs.

`property` Runs.more

Returns True if there are more runs to fetch. Returns False if there are no more runs to fetch.

`method` `Runs.convert_objects`

convert_objects()

Converts GraphQL edges to Runs objects.

`method` `Runs.histories`

histories(
    samples: int = 500,
    keys: Optional[List[str]] = None,
    x_axis: str = '_step',
    format: Literal['default', 'pandas', 'polars'] = 'default',
    stream: Literal['default', 'system'] = 'default'
)

Return sampled history metrics for all runs that fit the filters conditions.

Args:

samples: The number of samples to return per run
keys: Only return metrics for specific keys
x_axis: Use this metric as the xAxis defaults to _step
format: Format to return data in, options are “default”, “pandas”, “polars”
stream: “default” for metrics, “system” for machine metrics

Returns:

pandas.DataFrame: If format="pandas", returns a pandas.DataFrame of history metrics.
polars.DataFrame: If format="polars", returns a polars.DataFrame of history metrics.
list of dicts: If format="default", returns a list of dicts containing history metrics with a run_id key.

`class` `Run`

A single run associated with an entity and project.

Args:

client: The W&B API client.
entity: The entity associated with the run.
project: The project associated with the run.
run_id: The unique identifier for the run.
attrs: The attributes of the run.
include_sweeps: Whether to include sweeps in the run.

Attributes:

tags ([str]): a list of tags associated with the run
url (str): the url of this run
id (str): unique identifier for the run (defaults to eight characters)
name (str): the name of the run
state (str): one of: running, finished, crashed, killed, preempting, preempted
config (dict): a dict of hyperparameters associated with the run
created_at (str): ISO timestamp when the run was started
system_metrics (dict): the latest system metrics recorded for the run
summary (dict): A mutable dict-like property that holds the current summary. Calling update will persist any changes.
project (str): the project associated with the run
entity (str): the name of the entity associated with the run
project_internal_id (int): the internal id of the project
user (str): the name of the user who created the run
path (str): Unique identifier [entity]/[project]/[run_id]
notes (str): Notes about the run
read_only (boolean): Whether the run is editable
history_keys (str): Keys of the history metrics that have been logged
with wandb.log({key: value})
metadata (str): Metadata about the run from wandb-metadata.json

`method` `Run.init`

__init__(
    client: 'RetryingClient',
    entity: str,
    project: str,
    run_id: str,
    attrs: Optional[Mapping] = None,
    include_sweeps: bool = True
)

Initialize a Run object.

Run is always initialized by calling api.runs() where api is an instance of wandb.Api.

`property` Run.entity

The entity associated with the run.

`property` Run.id

The unique identifier for the run.

`property` Run.json_config

`property` Run.lastHistoryStep

Returns the last step logged in the run’s history.

`property` Run.metadata

Metadata about the run from wandb-metadata.json.

Metadata includes the run’s description, tags, start time, memory usage and more.

`property` Run.name

The name of the run.

`property` Run.path

The path of the run. The path is a list containing the entity, project, and run_id.

`property` Run.state

The state of the run. Can be one of: Finished, Failed, Crashed, or Running.

`property` Run.storage_id

The unique storage identifier for the run.

`property` Run.summary

A mutable dict-like property that holds summary values associated with the run.

`property` Run.url

The URL of the run.

The run URL is generated from the entity, project, and run_id. For SaaS users, it takes the form of https://wandb.ai/entity/project/run_id.

`property` Run.username

This API is deprecated. Use entity instead.

`classmethod` `Run.create`

create(api, run_id=None, project=None, entity=None)

Create a run for the given project.

`method` `Run.delete`

delete(delete_artifacts=False)

Delete the given run from the wandb backend.

Args:

delete_artifacts (bool, optional): Whether to delete the artifacts associated with the run.

`method` `Run.file`

file(name)

Return the path of a file with a given name in the artifact.

Args:

name (str): name of requested file.

Returns: A File matching the name argument.

`method` `Run.files`

files(names=None, per_page=50)

Return a file path for each file named.

Args:

names (list): names of the requested files, if empty returns all files
per_page (int): number of results per page.

Returns: A Files object, which is an iterator over File objects.

`method` `Run.history`

history(samples=500, keys=None, x_axis='_step', pandas=True, stream='default')

Return sampled history metrics for a run.

This is simpler and faster if you are ok with the history records being sampled.

Args:

samples : (int, optional) The number of samples to return
pandas : (bool, optional) Return a pandas dataframe
keys : (list, optional) Only return metrics for specific keys
x_axis : (str, optional) Use this metric as the xAxis defaults to _step
stream : (str, optional) “default” for metrics, “system” for machine metrics

Returns:

pandas.DataFrame: If pandas=True returns a pandas.DataFrame of history metrics.
list of dicts: If pandas=False returns a list of dicts of history metrics.

`method` `Run.load`

load(force=False)

Fetch and update run data from GraphQL database.

Ensures run data is up to date.

Args:

force (bool): Whether to force a refresh of the run data.

`method` `Run.log_artifact`

log_artifact(
    artifact: 'wandb.Artifact',
    aliases: Optional[Collection[str]] = None,
    tags: Optional[Collection[str]] = None
)

Declare an artifact as output of a run.

Args:

artifact (Artifact): An artifact returned from wandb.Api().artifact(name).
aliases (list, optional): Aliases to apply to this artifact.
tags: (list, optional) Tags to apply to this artifact, if any.

Returns: A Artifact object.

`method` `Run.logged_artifacts`

logged_artifacts(per_page: int = 100) → RunArtifacts

Fetches all artifacts logged by this run.

Retrieves all output artifacts that were logged during the run. Returns a paginated result that can be iterated over or collected into a single list.

Args:

per_page: Number of artifacts to fetch per API request.

Returns: An iterable collection of all Artifact objects logged as outputs during this run.

Example:

import wandb
import tempfile

with tempfile.NamedTemporaryFile(mode="w", delete=False, suffix=".txt") as tmp:
   tmp.write("This is a test artifact")
   tmp_path = tmp.name
run = wandb.init(project="artifact-example")
artifact = wandb.Artifact("test_artifact", type="dataset")
artifact.add_file(tmp_path)
run.log_artifact(artifact)
run.finish()

api = wandb.Api()

finished_run = api.run(f"{run.entity}/{run.project}/{run.id}")

for logged_artifact in finished_run.logged_artifacts():
   print(logged_artifact.name)

`method` `Run.save`

save()

Persist changes to the run object to the W&B backend.

`method` `Run.scan_history`

scan_history(keys=None, page_size=1000, min_step=None, max_step=None)

Returns an iterable collection of all history records for a run.

Args:

keys ([str], optional): only fetch these keys, and only fetch rows that have all of keys defined.
page_size (int, optional): size of pages to fetch from the api.
min_step (int, optional): the minimum number of pages to scan at a time.
max_step (int, optional): the maximum number of pages to scan at a time.

Returns: An iterable collection over history records (dict).

Example: Export all the loss values for an example run

run = api.run("entity/project-name/run-id")
history = run.scan_history(keys=["Loss"])
losses = [row["Loss"] for row in history]

`method` `Run.to_html`

to_html(height=420, hidden=False)

Generate HTML containing an iframe displaying this run.

`method` `Run.update`

update()

Persist changes to the run object to the wandb backend.

`method` `Run.upload_file`

upload_file(path, root='.')

Uploads a local file to W&B, associating it with this run.

Args:

path (str): Path to the file to upload. Can be absolute or relative.
root (str): The root path to save the file relative to. For example, if you want to have the file saved in the run as “my_dir/file.txt” and you’re currently in “my_dir” you would set root to “../”. Defaults to current directory (".").

Returns: A File object representing the uploaded file.

`method` `Run.use_artifact`

use_artifact(artifact, use_as=None)

Declare an artifact as an input to a run.

Args:

artifact (Artifact): An artifact returned from wandb.Api().artifact(name)
use_as (string, optional): A string identifying how the artifact is used in the script. Used to easily differentiate artifacts used in a run, when using the beta wandb launch feature’s artifact swapping functionality.

Returns: A Artifact object.

`method` `Run.used_artifacts`

used_artifacts(per_page: int = 100) → RunArtifacts

Fetches artifacts explicitly used by this run.

Retrieves only the input artifacts that were explicitly declared as used during the run, typically via run.use_artifact(). Returns a paginated result that can be iterated over or collected into a single list.

Args:

per_page: Number of artifacts to fetch per API request.

Returns: An iterable collection of Artifact objects explicitly used as inputs in this run.

Example:

import wandb

run = wandb.init(project="artifact-example")
run.use_artifact("test_artifact:latest")
run.finish()

api = wandb.Api()
finished_run = api.run(f"{run.entity}/{run.project}/{run.id}")
for used_artifact in finished_run.used_artifacts():
   print(used_artifact.name)
test_artifact

`method` `Run.wait_until_finished`

wait_until_finished()

Check the state of the run until it is finished.

10 - sweeps

View the source code

`module` `wandb.apis.public`

W&B Public API for Hyperparameter Sweeps.

This module provides classes for interacting with W&B hyperparameter optimization sweeps. Classes include:

Sweep: Represents a hyperparameter optimization sweep, providing access to:

Sweep configuration and state
Associated runs and their metrics
Best performing runs
URLs for visualization

Example:

from wandb.apis.public import Api

# Initialize API
api = Api()

# Get a specific sweep
sweep = api.sweep("entity/project/sweep_id")

# Access sweep properties
print(f"Sweep: {sweep.name}")
print(f"State: {sweep.state}")
print(f"Best Loss: {sweep.best_loss}")

# Get best performing run
best_run = sweep.best_run()
print(f"Best Run: {best_run.name}")
print(f"Metrics: {best_run.summary}")

Note:

This module is part of the W&B Public API and provides read-only access to sweep data. For creating and controlling sweeps, use the wandb.sweep() and wandb.agent() functions from the main wandb package.

`class` `Sweep`

The set of runs associated with the sweep.

Attributes:

runs (Runs): List of runs
id (str): Sweep ID
project (str): The name of the project the sweep belongs to
config (dict): Dictionary containing the sweep configuration
state (str): The state of the sweep. Can be “Finished”, “Failed”, “Crashed”, or “Running”.
expected_run_count (int): The number of expected runs for the sweep

`method` `Sweep.init`

__init__(client, entity, project, sweep_id, attrs=None)

`property` Sweep.config

The sweep configuration used for the sweep.

`property` Sweep.entity

The entity associated with the sweep.

`property` Sweep.expected_run_count

Return the number of expected runs in the sweep or None for infinite runs.

`property` Sweep.name

The name of the sweep.

If the sweep has a name, it will be returned. Otherwise, the sweep ID will be returned.

`property` Sweep.order

Return the order key for the sweep.

`property` Sweep.path

Returns the path of the project.

The path is a list containing the entity, project name, and sweep ID.

`property` Sweep.url

The URL of the sweep.

The sweep URL is generated from the entity, project, the term “sweeps”, and the sweep ID.run_id. For SaaS users, it takes the form of https://wandb.ai/entity/project/sweeps/sweeps_ID.

`property` Sweep.username

Note: Deprecated. Use entity instead.

`method` `Sweep.best_run`

best_run(order=None)

Return the best run sorted by the metric defined in config or the order passed in.

`classmethod` `Sweep.get`

get(
    client,
    entity=None,
    project=None,
    sid=None,
    order=None,
    query=None,
    **kwargs
)

Execute a query against the cloud backend.

`method` `Sweep.load`

load(force: bool = False)

Fetch and update sweep data logged to the run from GraphQL database.

`method` `Sweep.to_html`

to_html(height=420, hidden=False)

Generate HTML containing an iframe displaying this sweep.

11 - teams

View the source code

`module` `wandb.apis.public`

W&B Public API for managing teams and team members.

This module provides classes for managing W&B teams and their members. Classes include:

Team: Manage W&B teams and their settings

Create new teams
Invite team members
Create service accounts
Manage team permissions and settings

Member: Represent and manage team members

Access member information
Delete members
Manage member permissions

Note:

This module is part of the W&B Public API and provides methods to manage teams and their members. Team management operations require appropriate permissions.

`class` `Member`

A member of a team.

Args:

client (wandb.apis.internal.Api): The client instance to use
team (str): The name of the team this member belongs to
attrs (dict): The member attributes

`method` `Member.init`

__init__(client, team, attrs)

`method` `Member.delete`

delete()

Remove a member from a team.

Returns: Boolean indicating success

`class` `Team`

A class that represents a W&B team.

This class provides methods to manage W&B teams, including creating teams, inviting members, and managing service accounts. It inherits from Attrs to handle team attributes.

Args:

client (wandb.apis.public.Api): The api instance to use
name (str): The name of the team
attrs (dict): Optional dictionary of team attributes

Note:

Team management requires appropriate permissions.

`method` `Team.init`

__init__(client, name, attrs=None)

`classmethod` `Team.create`

create(api, team, admin_username=None)

Create a new team.

Args:

api: (Api) The api instance to use
team: (str) The name of the team
admin_username: (str) optional username of the admin user of the team, defaults to the current user.

Returns: A Team object

`method` `Team.create_service_account`

create_service_account(description)

Create a service account for the team.

Args:

description: (str) A description for this service account

Returns: The service account Member object, or None on failure

`method` `Team.invite`

invite(username_or_email, admin=False)

Invite a user to a team.

Args:

username_or_email: (str) The username or email address of the user you want to invite
admin: (bool) Whether to make this user a team admin, defaults to False

Returns: True on success, False if user was already invited or didn’t exist

`method` `Team.load`

load(force=False)

Return members that belong to a team.

12 - users

View the source code

`module` `wandb.apis.public`

W&B Public API for User Management.

This module provides classes for managing W&B users and their API keys. Classes include:

User: Manage W&B user accounts and authentication

Create new users
Generate and manage API keys
Access team memberships
Handle user properties and permissions

Note:

This module is part of the W&B Public API and provides methods to manage users and their authentication. Some operations require admin privileges.

`class` `User`

A class representing a W&B user with authentication and management capabilities.

This class provides methods to manage W&B users, including creating users, managing API keys, and accessing team memberships. It inherits from Attrs to handle user attributes.

Args:

client: (wandb.apis.internal.Api) The client instance to use
attrs: (dict) The user attributes

Note:

Some operations require admin privileges

`method` `User.init`

__init__(client, attrs)

`property` User.api_keys

List of API key names associated with the user.

Returns:

list[str]: Names of API keys associated with the user. Empty list if user has no API keys or if API key data hasn’t been loaded.

`property` User.teams

List of team names that the user is a member of.

Returns:

list (list): Names of teams the user belongs to. Empty list if user has no team memberships or if teams data hasn’t been loaded.

`property` User.user_api

An instance of the api using credentials from the user.

`classmethod` `User.create`

create(api, email, admin=False)

Create a new user.

Args:

api (Api): The api instance to use
email (str): The name of the team
admin (bool): Whether this user should be a global instance admin

Returns: A User object

`method` `User.delete_api_key`

delete_api_key(api_key)

Delete a user’s api key.

Args:

api_key (str): The name of the API key to delete. This should be one of the names returned by the api_keys property.

Returns: Boolean indicating success

Raises: ValueError if the api_key couldn’t be found

`method` `User.generate_api_key`

generate_api_key(description=None)

Generate a new api key.

Args:

description (str, optional): A description for the new API key. This can be used to identify the purpose of the API key.

Returns: The new api key, or None on failure

Analytics and Query API

1 - api

module wandb.apis.public

class RetryingClient

method RetryingClient.__init__

property RetryingClient.app_url

property RetryingClient.server_info

method RetryingClient.execute

method RetryingClient.version_supported

class Api

method Api.__init__

property Api.api_key

property Api.client

property Api.default_entity

property Api.user_agent

property Api.viewer

method Api.artifact

method Api.artifact_collection

method Api.artifact_collection_exists

method Api.artifact_collections

method Api.artifact_exists

method Api.artifact_type

method Api.artifact_types

method Api.artifact_versions

method Api.artifacts

method Api.create_project

method Api.create_run

method Api.create_run_queue

method Api.create_team

method Api.create_user

method Api.flush

method Api.from_path

method Api.list_jobs

method Api.project

method Api.projects

method Api.queued_run

method Api.registries

method Api.reports

method Api.run

method Api.run_queue

method Api.runs

method Api.sweep

method Api.sync_tensorboard

method Api.team

method Api.upsert_run_queue

method Api.user

method Api.users

2 - artifacts

module wandb.apis.public

function server_supports_artifact_collections_gql_edges

function artifact_collection_edge_name

function artifact_collection_plural_edge_name

class ArtifactTypes

method ArtifactTypes.__init__

property ArtifactTypes.cursor

property ArtifactTypes.length

property ArtifactTypes.more

method ArtifactTypes.convert_objects

method ArtifactTypes.update_variables

class ArtifactType

method ArtifactType.__init__

property ArtifactType.id

property ArtifactType.name

method ArtifactType.collection

method ArtifactType.collections

method ArtifactType.load

class ArtifactCollections

method ArtifactCollections.__init__

property ArtifactCollections.cursor

property ArtifactCollections.length

property ArtifactCollections.more

method ArtifactCollections.convert_objects

method ArtifactCollections.update_variables

class ArtifactCollection

method ArtifactCollection.__init__

property ArtifactCollection.aliases

property ArtifactCollection.created_at

property ArtifactCollection.description

property ArtifactCollection.id

property ArtifactCollection.name

`module` `wandb.apis.public`

`class` `RetryingClient`

`method` `RetryingClient.init`

`property` RetryingClient.app_url

`property` RetryingClient.server_info

`method` `RetryingClient.execute`

`method` `RetryingClient.version_supported`

`class` `Api`

`method` `Api.init`

`property` Api.api_key

`property` Api.client

`property` Api.default_entity

`property` Api.user_agent

`property` Api.viewer

`method` `Api.artifact`

`method` `Api.artifact_collection`

`method` `Api.artifact_collection_exists`

`method` `Api.artifact_collections`

`method` `Api.artifact_exists`

`method` `Api.artifact_type`

`method` `Api.artifact_types`

`method` `Api.artifact_versions`

`method` `Api.artifacts`

`method` `Api.create_project`

`method` `Api.create_run`

`method` `Api.create_run_queue`

`method` `Api.create_team`

`method` `Api.create_user`

`method` `Api.flush`

`method` `Api.from_path`

`method` `Api.list_jobs`

`method` `Api.project`

`method` `Api.projects`

`method` `Api.queued_run`

`method` `Api.registries`

`method` `Api.reports`

`method` `Api.run`

`method` `Api.run_queue`

`method` `Api.runs`

`method` `Api.sweep`

`method` `Api.sync_tensorboard`

`method` `Api.team`

`method` `Api.upsert_run_queue`

`method` `Api.user`

`method` `Api.users`

`module` `wandb.apis.public`

`function` `server_supports_artifact_collections_gql_edges`

`function` `artifact_collection_edge_name`

`function` `artifact_collection_plural_edge_name`

`class` `ArtifactTypes`

`method` `ArtifactTypes.init`

`property` ArtifactTypes.cursor

`property` ArtifactTypes.length

`property` ArtifactTypes.more

`method` `ArtifactTypes.convert_objects`

`method` `ArtifactTypes.update_variables`

`class` `ArtifactType`

`method` `ArtifactType.init`

`property` ArtifactType.id

`property` ArtifactType.name

`method` `ArtifactType.collection`

`method` `ArtifactType.collections`

`method` `ArtifactType.load`

`class` `ArtifactCollections`

`method` `ArtifactCollections.init`

`property` ArtifactCollections.cursor

`property` ArtifactCollections.length

`property` ArtifactCollections.more

`method` `ArtifactCollections.convert_objects`

`method` `ArtifactCollections.update_variables`

`class` `ArtifactCollection`

`method` `ArtifactCollection.init`

`property` ArtifactCollection.aliases

`property` ArtifactCollection.created_at

`property` ArtifactCollection.description

`property` ArtifactCollection.id

`property` ArtifactCollection.name

`property` ArtifactCollection.tags

`property` ArtifactCollection.type

`method` `ArtifactCollection.artifacts`