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

Return to the regular view of this page.

SDK

Train and fine-tune models, manage models from experimentation to production. For guides and examples, see https://docs.wandb.ai.

1 - Functions

1.1 - agent()

View the source code

function agent

agent(
    sweep_id: str,
    function: Optional[Callable] = None,
    entity: Optional[str] = None,
    project: Optional[str] = None,
    count: Optional[int] = None
) → None

Start one or more sweep agents.

The sweep agent uses the sweep_id to know which sweep it is a part of, what function to execute, and (optionally) how many agents to run.

Args:

  • sweep_id: The unique identifier for a sweep. A sweep ID is generated by W&B CLI or Python SDK.
  • function: A function to call instead of the “program” specified in the sweep config.
  • entity: The username or team name where you want to send W&B runs created by the sweep to. Ensure that the entity you specify already exists. If you don’t specify an entity, the run will be sent to your default entity, which is usually your username.
  • project: The name of the project where W&B runs created from the sweep are sent to. If the project is not specified, the run is sent to a project labeled “Uncategorized”.
  • count: The number of sweep config trials to try.

1.2 - controller()

View the source code

function controller

controller(
    sweep_id_or_config: Optional[str, Dict] = None,
    entity: Optional[str] = None,
    project: Optional[str] = None
) → _WandbController

Public sweep controller constructor.

Examples:

import wandb

tuner = wandb.controller(...)
print(tuner.sweep_config)
print(tuner.sweep_id)
tuner.configure_search(...)
tuner.configure_stopping(...)

1.3 - finish()

View the source code

function finish

finish(exit_code: 'int | None' = None, quiet: 'bool | None' = None) → None

Finish a run and upload any remaining data.

Marks the completion of a W&B run and ensures all data is synced to the server. The run’s final state is determined by its exit conditions and sync status.

Run States:

  • Running: Active run that is logging data and/or sending heartbeats.
  • Crashed: Run that stopped sending heartbeats unexpectedly.
  • Finished: Run completed successfully (exit_code=0) with all data synced.
  • Failed: Run completed with errors (exit_code!=0).

Args:

  • exit_code: Integer indicating the run’s exit status. Use 0 for success, any other value marks the run as failed.
  • quiet: Deprecated. Configure logging verbosity using wandb.Settings(quiet=...).

1.4 - init()

View the source code

function init

init(
    entity: 'str | None' = None,
    project: 'str | None' = None,
    dir: 'StrPath | None' = None,
    id: 'str | None' = None,
    name: 'str | None' = None,
    notes: 'str | None' = None,
    tags: 'Sequence[str] | None' = None,
    config: 'dict[str, Any] | str | None' = None,
    config_exclude_keys: 'list[str] | None' = None,
    config_include_keys: 'list[str] | None' = None,
    allow_val_change: 'bool | None' = None,
    group: 'str | None' = None,
    job_type: 'str | None' = None,
    mode: "Literal['online', 'offline', 'disabled'] | None" = None,
    force: 'bool | None' = None,
    anonymous: "Literal['never', 'allow', 'must'] | None" = None,
    reinit: "bool | Literal[None, 'default', 'return_previous', 'finish_previous', 'create_new']" = None,
    resume: "bool | Literal['allow', 'never', 'must', 'auto'] | None" = None,
    resume_from: 'str | None' = None,
    fork_from: 'str | None' = None,
    save_code: 'bool | None' = None,
    tensorboard: 'bool | None' = None,
    sync_tensorboard: 'bool | None' = None,
    monitor_gym: 'bool | None' = None,
    settings: 'Settings | dict[str, Any] | None' = None
) → Run

Start a new run to track and log to W&B.

In an ML training pipeline, you could add wandb.init() to the beginning of your training script as well as your evaluation script, and each piece would be tracked as a run in W&B.

wandb.init() spawns a new background process to log data to a run, and it also syncs data to https://wandb.ai by default, so you can see your results in real-time. When you’re done logging data, call wandb.finish() to end the run. If you don’t call run.finish(), the run will end when your script exits.

Run IDs must not contain any of the following special characters / \ # ? % :

Args:

  • entity: The username or team name the runs are logged to. The entity must already exist, so ensure you create your account or team in the UI before starting to log runs. If not specified, the run will default your default entity. To change the default entity, go to your settings and update the “Default location to create new projects” under “Default team”.
  • project: The name of the project under which this run will be logged. If not specified, we use a heuristic to infer the project name based on the system, such as checking the git root or the current program file. If we can’t infer the project name, the project will default to "uncategorized".
  • dir: The absolute path to the directory where experiment logs and metadata files are stored. If not specified, this defaults to the ./wandb directory. Note that this does not affect the location where artifacts are stored when calling download().
  • id: A unique identifier for this run, used for resuming. It must be unique within the project and cannot be reused once a run is deleted. For a short descriptive name, use the name field, or for saving hyperparameters to compare across runs, use config.
  • name: A short display name for this run, which appears in the UI to help you identify it. By default, we generate a random two-word name allowing easy cross-reference runs from table to charts. Keeping these run names brief enhances readability in chart legends and tables. For saving hyperparameters, we recommend using the config field.
  • notes: A detailed description of the run, similar to a commit message in Git. Use this argument to capture any context or details that may help you recall the purpose or setup of this run in the future.
  • tags: A list of tags to label this run in the UI. Tags are helpful for organizing runs or adding temporary identifiers like “baseline” or “production.” You can easily add, remove tags, or filter by tags in the UI. If resuming a run, the tags provided here will replace any existing tags. To add tags to a resumed run without overwriting the current tags, use run.tags += ("new_tag",) after calling run = wandb.init().
  • config: Sets wandb.config, a dictionary-like object for storing input parameters to your run, such as model hyperparameters or data preprocessing settings. The config appears in the UI in an overview page, allowing you to group, filter, and sort runs based on these parameters. Keys should not contain periods (.), and values should be smaller than 10 MB. If a dictionary, argparse.Namespace, or absl.flags.FLAGS is provided, the key-value pairs will be loaded directly into wandb.config. If a string is provided, it is interpreted as a path to a YAML file, from which configuration values will be loaded into wandb.config.
  • config_exclude_keys: A list of specific keys to exclude from wandb.config.
  • config_include_keys: A list of specific keys to include in wandb.config.
  • allow_val_change: Controls whether config values can be modified after their initial set. By default, an exception is raised if a config value is overwritten. For tracking variables that change during training, such as a learning rate, consider using wandb.log() instead. By default, this is False in scripts and True in Notebook environments.
  • group: Specify a group name to organize individual runs as part of a larger experiment. This is useful for cases like cross-validation or running multiple jobs that train and evaluate a model on different test sets. Grouping allows you to manage related runs collectively in the UI, making it easy to toggle and review results as a unified experiment.
  • job_type: Specify the type of run, especially helpful when organizing runs within a group as part of a larger experiment. For example, in a group, you might label runs with job types such as “train” and “eval”. Defining job types enables you to easily filter and group similar runs in the UI, facilitating direct comparisons.
  • mode: Specifies how run data is managed, with the following options:
    • "online" (default): Enables live syncing with W&B when a network connection is available, with real-time updates to visualizations.
    • "offline": Suitable for air-gapped or offline environments; data is saved locally and can be synced later. Ensure the run folder is preserved to enable future syncing.
    • "disabled": Disables all W&B functionality, making the run’s methods no-ops. Typically used in testing to bypass W&B operations.
  • force: Determines if a W&B login is required to run the script. If True, the user must be logged in to W&B; otherwise, the script will not proceed. If False (default), the script can proceed without a login, switching to offline mode if the user is not logged in.
  • anonymous: Specifies the level of control over anonymous data logging. Available options are:
    • "never" (default): Requires you to link your W&B account before tracking the run. This prevents unintentional creation of anonymous runs by ensuring each run is associated with an account.
    • "allow": Enables a logged-in user to track runs with their account, but also allows someone running the script without a W&B account to view the charts and data in the UI.
    • "must": Forces the run to be logged to an anonymous account, even if the user is logged in.
  • reinit: Shorthand for the “reinit” setting. Determines the behavior of wandb.init() when a run is active.
  • resume: Controls the behavior when resuming a run with the specified id. Available options are:
    • "allow": If a run with the specified id exists, it will resume from the last step; otherwise, a new run will be created.
    • "never": If a run with the specified id exists, an error will be raised. If no such run is found, a new run will be created.
    • "must": If a run with the specified id exists, it will resume from the last step. If no run is found, an error will be raised.
    • "auto": Automatically resumes the previous run if it crashed on this machine; otherwise, starts a new run.
    • True: Deprecated. Use "auto" instead.
    • False: Deprecated. Use the default behavior (leaving resume unset) to always start a new run. If resume is set, fork_from and resume_from cannot be used. When resume is unset, the system will always start a new run.
  • resume_from: Specifies a moment in a previous run to resume a run from, using the format {run_id}?_step={step}. This allows users to truncate the history logged to a run at an intermediate step and resume logging from that step. The target run must be in the same project. If an id argument is also provided, the resume_from argument will take precedence. resume, resume_from and fork_from cannot be used together, only one of them can be used at a time. Note that this feature is in beta and may change in the future.
  • fork_from: Specifies a point in a previous run from which to fork a new run, using the format {id}?_step={step}. This creates a new run that resumes logging from the specified step in the target run’s history. The target run must be part of the current project. If an id argument is also provided, it must be different from the fork_from argument, an error will be raised if they are the same. resume, resume_from and fork_from cannot be used together, only one of them can be used at a time. Note that this feature is in beta and may change in the future.
  • save_code: Enables saving the main script or notebook to W&B, aiding in experiment reproducibility and allowing code comparisons across runs in the UI. By default, this is disabled, but you can change the default to enable on your settings page.
  • tensorboard: Deprecated. Use sync_tensorboard instead.
  • sync_tensorboard: Enables automatic syncing of W&B logs from TensorBoard or TensorBoardX, saving relevant event files for viewing in the W&B UI.
  • saving relevant event files for viewing in the W&B UI. (Default: False)
  • monitor_gym: Enables automatic logging of videos of the environment when using OpenAI Gym.
  • settings: Specifies a dictionary or wandb.Settings object with advanced settings for the run.

Raises:

  • Error: If some unknown or internal error happened during the run initialization.
  • AuthenticationError: If the user failed to provide valid credentials.
  • CommError: If there was a problem communicating with the WandB server.
  • UsageError: If the user provided invalid arguments.
  • KeyboardInterrupt: If user interrupts the run.

Returns: A Run object.

Examples: wandb.init() returns a Run object. Use the run object to log data, save artifacts, and manage the run lifecycle.

import wandb

config = {"lr": 0.01, "batch_size": 32}
with wandb.init(config=config) as run:
    # Log accuracy and loss to the run
    acc = 0.95  # Example accuracy
    loss = 0.05  # Example loss
    run.log({"accuracy": acc, "loss": loss})

1.5 - login()

View the source code

function login

login(
    anonymous: Optional[Literal['allow', 'must', 'never']] = None,
    key: Optional[str] = None,
    relogin: Optional[bool] = None,
    host: Optional[str] = None,
    force: Optional[bool] = None,
    timeout: Optional[int] = None,
    verify: bool = False,
    referrer: Optional[str] = None
) → bool

Set up W&B login credentials.

By default, this will only store credentials locally without verifying them with the W&B server. To verify credentials, pass verify=True.

Args:

  • anonymous: Set to “must”, “allow”, or “never”. If set to “must”, always log a user in anonymously. If set to “allow”, only create an anonymous user if the user isn’t already logged in. If set to “never”, never log a user anonymously. Default set to “never”.
  • key: The API key to use.
  • relogin: If true, will re-prompt for API key.
  • host: The host to connect to.
  • force: If true, will force a relogin.
  • timeout: Number of seconds to wait for user input.
  • verify: Verify the credentials with the W&B server.
  • referrer: The referrer to use in the URL login request.

Returns:

  • bool: If key is configured.

Raises:

  • AuthenticationError: If api_key fails verification with the server.
  • UsageError: If api_key cannot be configured and no tty.

1.6 - restore()

View the source code

function restore

restore(
    name: 'str',
    run_path: 'str | None' = None,
    replace: 'bool' = False,
    root: 'str | None' = None
) → None | TextIO

Download the specified file from cloud storage.

File is placed into the current directory or run directory. By default, will only download the file if it doesn’t already exist.

Args:

  • name: The name of the file.
  • run_path: Optional path to a run to pull files from, i.e. username/project_name/run_id if wandb.init has not been called, this is required.
  • replace: Whether to download the file even if it already exists locally
  • root: The directory to download the file to. Defaults to the current directory or the run directory if wandb.init was called.

Returns: None if it can’t find the file, otherwise a file object open for reading.

Raises:

  • CommError: If W&B can’t connect to the W&B backend.
  • ValueError: If the file is not found or can’t find run_path.

1.7 - setup()

View the source code

function setup

setup(settings: 'Settings | None' = None) → _WandbSetup

Prepares W&B for use in the current process and its children.

You can usually ignore this as it is implicitly called by wandb.init().

When using wandb in multiple processes, calling wandb.setup() in the parent process before starting child processes may improve performance and resource utilization.

Note that wandb.setup() modifies os.environ, and it is important that child processes inherit the modified environment variables.

See also wandb.teardown().

Args:

  • settings: Configuration settings to apply globally. These can be overridden by subsequent wandb.init() calls.

Example:

import multiprocessing

import wandb


def run_experiment(params):
   with wandb.init(config=params):
        # Run experiment
        pass


if __name__ == "__main__":
   # Start backend and set global config
   wandb.setup(settings={"project": "my_project"})

   # Define experiment parameters
   experiment_params = [
        {"learning_rate": 0.01, "epochs": 10},
        {"learning_rate": 0.001, "epochs": 20},
   ]

   # Start multiple processes, each running a separate experiment
   processes = []
   for params in experiment_params:
        p = multiprocessing.Process(target=run_experiment, args=(params,))
        p.start()
        processes.append(p)

   # Wait for all processes to complete
   for p in processes:
        p.join()

   # Optional: Explicitly shut down the backend
   wandb.teardown()

1.8 - sweep()

View the source code

function sweep

sweep(
    sweep: Union[dict, Callable],
    entity: Optional[str] = None,
    project: Optional[str] = None,
    prior_runs: Optional[List[str]] = None
) → str

Initialize a hyperparameter sweep.

Search for hyperparameters that optimizes a cost function of a machine learning model by testing various combinations.

Make note the unique identifier, sweep_id, that is returned. At a later step provide the sweep_id to a sweep agent.

See Sweep configuration structure for information on how to define your sweep.

Args:

  • sweep: The configuration of a hyperparameter search. (or configuration generator). If you provide a callable, ensure that the callable does not take arguments and that it returns a dictionary that conforms to the W&B sweep config spec.
  • entity: The username or team name where you want to send W&B runs created by the sweep to. Ensure that the entity you specify already exists. If you don’t specify an entity, the run will be sent to your default entity, which is usually your username.
  • project: The name of the project where W&B runs created from the sweep are sent to. If the project is not specified, the run is sent to a project labeled ‘Uncategorized’.
  • prior_runs: The run IDs of existing runs to add to this sweep.

Returns:

  • sweep_id: (str) A unique identifier for the sweep.

1.9 - teardown()

View the source code

function teardown

teardown(exit_code: 'int | None' = None) → None

Waits for W&B to finish and frees resources.

Completes any runs that were not explicitly finished using run.finish() and waits for all data to be uploaded.

It is recommended to call this at the end of a session that used wandb.setup(). It is invoked automatically in an atexit hook, but this is not reliable in certain setups such as when using Python’s multiprocessing module.

2 - Classes

2.1 - Artifact

View the source code

class Artifact

Flexible and lightweight building block for dataset and model versioning.

Construct an empty W&B Artifact. Populate an artifacts contents with methods that begin with add. Once the artifact has all the desired files, you can call run.log_artifact() to log it.

Args:

  • name (str): A human-readable name for the artifact. Use the name to identify a specific artifact in the W&B App UI or programmatically. You can interactively reference an artifact with the use_artifact Public API. A name can contain letters, numbers, underscores, hyphens, and dots. The name must be unique across a project.
  • type (str): The artifact’s type. Use the type of an artifact to both organize and differentiate artifacts. You can use any string that contains letters, numbers, underscores, hyphens, and dots. Common types include dataset or model. Include model within your type string if you want to link the artifact to the W&B Model Registry. Note that some types reserved for internal use and cannot be set by users. Such types include job and types that start with wandb-.
  • description (str | None) = None: A description of the artifact. For Model or Dataset Artifacts, add documentation for your standardized team model or dataset card. View an artifact’s description programmatically with the Artifact.description attribute or programmatically with the W&B App UI. W&B renders the description as markdown in the W&B App.
  • metadata (dict[str, Any] | None) = None: Additional information about an artifact. Specify metadata as a dictionary of key-value pairs. You can specify no more than 100 total keys.
  • incremental: Use Artifact.new_draft() method instead to modify an existing artifact.
  • use_as: Deprecated.
  • is_link: Boolean indication of if the artifact is a linked artifact(True) or source artifact(False).

Returns: An Artifact object.

method Artifact.__init__

__init__(
    name: 'str',
    type: 'str',
    description: 'str | None' = None,
    metadata: 'dict[str, Any] | None' = None,
    incremental: 'bool' = False,
    use_as: 'str | None' = None
) → None

property Artifact.aliases

List of one or more semantically-friendly references or

identifying “nicknames” assigned to an artifact version.

Aliases are mutable references that you can programmatically reference. Change an artifact’s alias with the W&B App UI or programmatically. See Create new artifact versions for more information.

property Artifact.collection

The collection this artifact was retrieved from.

A collection is an ordered group of artifact versions. If this artifact was retrieved from a portfolio / linked collection, that collection will be returned rather than the collection that an artifact version originated from. The collection that an artifact originates from is known as the source sequence.

property Artifact.commit_hash

The hash returned when this artifact was committed.

property Artifact.created_at

Timestamp when the artifact was created.

property Artifact.description

A description of the artifact.

property Artifact.digest

The logical digest of the artifact.

The digest is the checksum of the artifact’s contents. If an artifact has the same digest as the current latest version, then log_artifact is a no-op.

property Artifact.entity

The name of the entity that the artifact collection belongs to.

If the artifact is a link, the entity will be the entity of the linked artifact.

property Artifact.file_count

The number of files (including references).

property Artifact.history_step

The nearest step at which history metrics were logged for the source run of the artifact.

Examples:

run = artifact.logged_by()
if run and (artifact.history_step is not None):
    history = run.sample_history(
        min_step=artifact.history_step,
        max_step=artifact.history_step + 1,
        keys=["my_metric"],
    )

property Artifact.id

The artifact’s ID.

Boolean flag indicating if the artifact is a link artifact.

True: The artifact is a link artifact to a source artifact. False: The artifact is a source artifact.

property Artifact.linked_artifacts

Returns a list of all the linked artifacts of a source artifact.

If the artifact is a link artifact (artifact.is_link == True), it will return an empty list. Limited to 500 results.

property Artifact.manifest

The artifact’s manifest.

The manifest lists all of its contents, and can’t be changed once the artifact has been logged.

property Artifact.metadata

User-defined artifact metadata.

Structured data associated with the artifact.

property Artifact.name

The artifact name and version of the artifact.

A string with the format {collection}:{alias}. If fetched before an artifact is logged/saved, the name won’t contain the alias. If the artifact is a link, the name will be the name of the linked artifact.

property Artifact.project

The name of the project that the artifact collection belongs to.

If the artifact is a link, the project will be the project of the linked artifact.

property Artifact.qualified_name

The entity/project/name of the artifact.

If the artifact is a link, the qualified name will be the qualified name of the linked artifact path.

property Artifact.size

The total size of the artifact in bytes.

Includes any references tracked by this artifact.

property Artifact.source_artifact

Returns the source artifact. The source artifact is the original logged artifact.

If the artifact itself is a source artifact (artifact.is_link == False), it will return itself.

property Artifact.source_collection

The artifact’s source collection.

The source collection is the collection that the artifact was logged from.

property Artifact.source_entity

The name of the entity of the source artifact.

property Artifact.source_name

The artifact name and version of the source artifact.

A string with the format {source_collection}:{alias}. Before the artifact is saved, contains only the name since the version is not yet known.

property Artifact.source_project

The name of the project of the source artifact.

property Artifact.source_qualified_name

The source_entity/source_project/source_name of the source artifact.

property Artifact.source_version

The source artifact’s version.

A string with the format v{number}.

property Artifact.state

The status of the artifact. One of: “PENDING”, “COMMITTED”, or “DELETED”.

property Artifact.tags

List of one or more tags assigned to this artifact version.

property Artifact.ttl

The time-to-live (TTL) policy of an artifact.

Artifacts are deleted shortly after a TTL policy’s duration passes. If set to None, the artifact deactivates TTL policies and will be not scheduled for deletion, even if there is a team default TTL. An artifact inherits a TTL policy from the team default if the team administrator defines a default TTL and there is no custom policy set on an artifact.

Raises:

  • ArtifactNotLoggedError: Unable to fetch inherited TTL if the artifact has not been logged or saved.

property Artifact.type

The artifact’s type. Common types include dataset or model.

property Artifact.updated_at

The time when the artifact was last updated.

property Artifact.url

Constructs the URL of the artifact.

Returns:

  • str: The URL of the artifact.

property Artifact.use_as

Deprecated.

property Artifact.version

The artifact’s version.

A string with the format v{number}. If the artifact is a link artifact, the version will be from the linked collection.

method Artifact.add

add(
    obj: 'WBValue',
    name: 'StrPath',
    overwrite: 'bool' = False
) → ArtifactManifestEntry

Add wandb.WBValue obj to the artifact.

Args:

  • obj: The object to add. Currently support one of Bokeh, JoinedTable, PartitionedTable, Table, Classes, ImageMask, BoundingBoxes2D, Audio, Image, Video, Html, Object3D
  • name: The path within the artifact to add the object.
  • overwrite: If True, overwrite existing objects with the same file path if applicable.

Returns: The added manifest entry

Raises:

  • ArtifactFinalizedError: You cannot make changes to the current artifact version because it is finalized. Log a new artifact version instead.

method Artifact.add_dir

add_dir(
    local_path: 'str',
    name: 'str | None' = None,
    skip_cache: 'bool | None' = False,
    policy: "Literal['mutable', 'immutable'] | None" = 'mutable',
    merge: 'bool' = False
) → None

Add a local directory to the artifact.

Args:

  • local_path: The path of the local directory.
  • name: The subdirectory name within an artifact. The name you specify appears in the W&B App UI nested by artifact’s type. Defaults to the root of the artifact.
  • skip_cache: If set to True, W&B will not copy/move files to the cache while uploading
  • policy: By default, “mutable”.
    • mutable: Create a temporary copy of the file to prevent corruption during upload.
    • immutable: Disable protection, rely on the user not to delete or change the file.
  • merge: If False (default), throws ValueError if a file was already added in a previous add_dir call and its content has changed. If True, overwrites existing files with changed content. Always adds new files and never removes files. To replace an entire directory, pass a name when adding the directory using add_dir(local_path, name=my_prefix) and call remove(my_prefix) to remove the directory, then add it again.

Raises:

  • ArtifactFinalizedError: You cannot make changes to the current artifact version because it is finalized. Log a new artifact version instead.
  • ValueError: Policy must be “mutable” or “immutable”

method Artifact.add_file

add_file(
    local_path: 'str',
    name: 'str | None' = None,
    is_tmp: 'bool | None' = False,
    skip_cache: 'bool | None' = False,
    policy: "Literal['mutable', 'immutable'] | None" = 'mutable',
    overwrite: 'bool' = False
) → ArtifactManifestEntry

Add a local file to the artifact.

Args:

  • local_path: The path to the file being added.
  • name: The path within the artifact to use for the file being added. Defaults to the basename of the file.
  • is_tmp: If true, then the file is renamed deterministically to avoid collisions.
  • skip_cache: If True, do not copy files to the cache after uploading.
  • policy: By default, set to “mutable”. If set to “mutable”, create a temporary copy of the file to prevent corruption during upload. If set to “immutable”, disable protection and rely on the user not to delete or change the file.
  • overwrite: If True, overwrite the file if it already exists.

Returns: The added manifest entry.

Raises:

  • ArtifactFinalizedError: You cannot make changes to the current artifact version because it is finalized. Log a new artifact version instead.
  • ValueError: Policy must be “mutable” or “immutable”

method Artifact.add_reference

add_reference(
    uri: 'ArtifactManifestEntry | str',
    name: 'StrPath | None' = None,
    checksum: 'bool' = True,
    max_objects: 'int | None' = None
) → Sequence[ArtifactManifestEntry]

Add a reference denoted by a URI to the artifact.

Unlike files or directories that you add to an artifact, references are not uploaded to W&B. For more information, see Track external files.

By default, the following schemes are supported:

  • 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 are pulled from the object metadata. If bucket versioning is enabled, then the version ID is also tracked.
  • gs: The checksum and size are pulled from the object metadata. If bucket versioning is enabled, then the version ID is also tracked.
  • https, domain matching *.blob.core.windows.net
  • Azure: The checksum and size are be pulled from the blob metadata. If storage account versioning is enabled, then the version ID is also tracked.
  • file: The checksum and size are 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.

Args:

  • uri: The URI path of the reference to add. The URI path can be an object returned from Artifact.get_entry to store a reference to another artifact’s entry.
  • name: The path within the artifact to place the contents of this reference.
  • checksum: Whether or not to checksum the resource(s) located at the reference URI. Checksumming is strongly recommended as it enables automatic integrity validation. Disabling checksumming will speed up artifact creation but reference directories will not iterated through so the objects in the directory will not be saved to the artifact. We recommend setting checksum=False when adding reference objects, in which case a new version will only be created if the reference URI changes.
  • max_objects: The maximum number of objects to consider when adding a reference that points to directory or bucket store prefix. By default, the maximum number of objects allowed for Amazon S3, GCS, Azure, and local files is 10,000,000. Other URI schemas do not have a maximum.

Returns: The added manifest entries.

Raises:

  • ArtifactFinalizedError: You cannot make changes to the current artifact version because it is finalized. Log a new artifact version instead.

method Artifact.checkout

checkout(root: 'str | None' = None) → str

Replace 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.

Args:

  • root: The directory to replace with this artifact’s files.

Returns: The path of the checked out contents.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.delete

delete(delete_aliases: 'bool' = False) → None

Delete an artifact and its files.

If called on a linked artifact, only the link is deleted, and the source artifact is unaffected.

Use artifact.unlink() instead of artifact.delete() to remove a link between a source artifact and a linked artifact.

Args:

  • delete_aliases: If set to True, deletes all aliases associated with the artifact. Otherwise, this raises an exception if the artifact has existing aliases. This parameter is ignored if the artifact is linked (a member of a portfolio collection).

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.download

download(
    root: 'StrPath | None' = None,
    allow_missing_references: 'bool' = False,
    skip_cache: 'bool | None' = None,
    path_prefix: 'StrPath | None' = None,
    multipart: 'bool | None' = None
) → FilePathStr

Download the contents of the artifact to the specified root directory.

Existing files located within root are not modified. Explicitly delete root before you call download if you want the contents of root to exactly match the artifact.

Args:

  • root: The directory W&B stores the artifact’s files.
  • allow_missing_references: If set to True, any invalid reference paths will be ignored while downloading referenced files.
  • skip_cache: If set to True, the artifact cache will be skipped when downloading and W&B will download each file into the default root or specified download directory.
  • path_prefix: If specified, only files with a path that starts with the given prefix will be downloaded. Uses unix format (forward slashes).
  • multipart: If set to None (default), the artifact will be downloaded in parallel using multipart download if individual file size is greater than 2GB. If set to True or False, the artifact will be downloaded in parallel or serially regardless of the file size.

Returns: The path to the downloaded contents.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.file

file(root: 'str | None' = None) → StrPath

Download a single file artifact to the directory you specify with root.

Args:

  • root: The root directory to store the file. Defaults to ./artifacts/self.name/.

Returns: The full path of the downloaded file.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.
  • ValueError: If the artifact contains more than one file.

method Artifact.files

files(names: 'list[str] | None' = None, per_page: 'int' = 50) → ArtifactFiles

Iterate over all files stored in this artifact.

Args:

  • names: The filename paths relative to the root of the artifact you wish to list.
  • per_page: The number of files to return per request.

Returns: An iterator containing File objects.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.finalize

finalize() → None

Finalize the artifact version.

You cannot modify an artifact version once it is finalized because the artifact is logged as a specific artifact version. Create a new artifact version to log more data to an artifact. An artifact is automatically finalized when you log the artifact with log_artifact.

method Artifact.get

get(name: 'str') → WBValue | None

Get the WBValue object located at the artifact relative name.

Args:

  • name: The artifact relative name to retrieve.

Returns: W&B object that can be logged with run.log() and visualized in the W&B UI.

Raises:

  • ArtifactNotLoggedError: if the artifact isn’t logged or the run is offline.

method Artifact.get_added_local_path_name

get_added_local_path_name(local_path: 'str') → str | None

Get the artifact relative name of a file added by a local filesystem path.

Args:

  • local_path: The local path to resolve into an artifact relative name.

Returns: The artifact relative name.

method Artifact.get_entry

get_entry(name: 'StrPath') → ArtifactManifestEntry

Get the entry with the given name.

Args:

  • name: The artifact relative name to get

Returns: A W&B object.

Raises:

  • ArtifactNotLoggedError: if the artifact isn’t logged or the run is offline.
  • KeyError: if the artifact doesn’t contain an entry with the given name.

method Artifact.get_path

get_path(name: 'StrPath') → ArtifactManifestEntry

Deprecated. Use get_entry(name).

method Artifact.is_draft

is_draft() → bool

Check if artifact is not saved.

Returns: Boolean. False if artifact is saved. True if artifact is not saved.

method Artifact.json_encode

json_encode() → dict[str, Any]

Returns the artifact encoded to the JSON format.

Returns: A dict with string keys representing attributes of the artifact.

link(target_path: 'str', aliases: 'list[str] | None' = None) → Artifact | None

Link this artifact to a portfolio (a promoted collection of artifacts).

Args:

  • target_path: The path to the portfolio inside a project. The target path must adhere to one of the following schemas {portfolio}, {project}/{portfolio} or {entity}/{project}/{portfolio}. To link the artifact to the Model Registry, rather than to a generic portfolio inside a project, set target_path to the following schema {"model-registry"}/{Registered Model Name} or {entity}/{"model-registry"}/{Registered Model Name}.
  • aliases: A list of strings that uniquely identifies the artifact inside the specified portfolio.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

Returns: The linked artifact if linking was successful, otherwise None.

method Artifact.logged_by

logged_by() → Run | None

Get the W&B run that originally logged the artifact.

Returns: The name of the W&B run that originally logged the artifact.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.new_draft

new_draft() → Artifact

Create a new draft artifact with the same content as this committed artifact.

Modifying an existing artifact creates a new artifact version known as an “incremental artifact”. The artifact returned can be extended or modified and logged as a new version.

Returns: An Artifact object.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.new_file

new_file(
    name: 'str',
    mode: 'str' = 'x',
    encoding: 'str | None' = None
) → Iterator[IO]

Open a new temporary file and add it to the artifact.

Args:

  • name: The name of the new file to add to the artifact.
  • mode: The file access mode to use to open the new file.
  • encoding: The encoding used to open the new file.

Returns: A new file object that can be written to. Upon closing, the file is automatically added to the artifact.

Raises:

  • ArtifactFinalizedError: You cannot make changes to the current artifact version because it is finalized. Log a new artifact version instead.

method Artifact.remove

remove(item: 'StrPath | ArtifactManifestEntry') → None

Remove an item from the artifact.

Args:

  • item: The item to remove. Can be a specific manifest entry or the name of an artifact-relative path. If the item matches a directory all items in that directory will be removed.

Raises:

  • ArtifactFinalizedError: You cannot make changes to the current artifact version because it is finalized. Log a new artifact version instead.
  • FileNotFoundError: If the item isn’t found in the artifact.

method Artifact.save

save(
    project: 'str | None' = None,
    settings: 'wandb.Settings | None' = None
) → None

Persist 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” is created to track this artifact.

Args:

  • project: A project to use for the artifact in the case that a run is not already in context.
  • settings: A settings object to use when initializing an automatic run. Most commonly used in testing harness.
unlink() → None

Unlink this artifact if it is currently a member of a promoted collection of artifacts.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.
  • ValueError: If the artifact is not linked, in other words, it is not a member of a portfolio collection.

method Artifact.used_by

used_by() → list[Run]

Get a list of the runs that have used this artifact and its linked artifacts.

Returns: A list of Run objects.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.

method Artifact.verify

verify(root: 'str | None' = None) → None

Verify that the contents of an artifact match the manifest.

All files in the directory are checksummed and the checksums are then cross-referenced against the artifact’s manifest. References are not verified.

Args:

  • root: The directory to verify. If None artifact will be downloaded to ‘./artifacts/self.name/’.

Raises:

  • ArtifactNotLoggedError: If the artifact is not logged.
  • ValueError: If the verification fails.

method Artifact.wait

wait(timeout: 'int | None' = None) → Artifact

If needed, wait for this artifact to finish logging.

Args:

  • timeout: The time, in seconds, to wait.

Returns: An Artifact object.

2.2 - Run

View the source code

class Run

A unit of computation logged by W&B. Typically, this is an ML experiment.

Call wandb.init() to create a new run. wandb.init() starts a new run and returns a wandb.Run object. Each run is associated with a unique ID (run ID). W&B recommends using a context (with statement) manager to automatically finish the run.

For distributed training experiments, you can either track each process separately using one run per process or track all processes to a single run. See Log distributed training experiments for more information.

You can log data to a run with wandb.Run.log(). Anything you log using wandb.Run.log() is sent to that run. See Create an experiment or wandb.init API reference page or more information.

There is a another Run object in the wandb.apis.public namespace. Use this object is to interact with runs that have already been created.

Attributes:

  • summary: (Summary) A summary of the run, which is a dictionary-like object. For more information, see
  • [Log summary metrics](https: //docs.wandb.ai/guides/track/log/log-summary/).

Examples: Create a run with wandb.init():

import wandb

# Start a new run and log some data
# Use context manager (`with` statement) to automatically finish the run
with wandb.init(entity="entity", project="project") as run:
    run.log({"accuracy": acc, "loss": loss})

property Run.config

Config object associated with this run.

property Run.config_static

Static config object associated with this run.

property Run.dir

The directory where files associated with the run are saved.

property Run.disabled

True if the run is disabled, False otherwise.

property Run.entity

The name of the W&B entity associated with the run.

Entity can be a username or the name of a team or organization.

property Run.group

Returns the name of the group associated with this run.

Grouping runs together allows related experiments to be organized and visualized collectively in the W&B UI. This is especially useful for scenarios such as distributed training or cross-validation, where multiple runs should be viewed and managed as a unified experiment.

In shared mode, where all processes share the same run object, setting a group is usually unnecessary, since there is only one run and no grouping is required.

property Run.id

Identifier for this run.

property Run.job_type

Name of the job type associated with the run.

View a run’s job type in the run’s Overview page in the W&B App.

You can use this to categorize runs by their job type, such as “training”, “evaluation”, or “inference”. This is useful for organizing and filtering runs in the W&B UI, especially when you have multiple runs with different job types in the same project. For more information, see Organize runs.

property Run.name

Display name of the run.

Display names are not guaranteed to be unique and may be descriptive. By default, they are randomly generated.

property Run.notes

Notes associated with the run, if there are any.

Notes can be a multiline string and can also use markdown and latex equations inside $$, like $x + 3$.

property Run.offline

True if the run is offline, False otherwise.

property Run.path

Path to the run.

Run paths include entity, project, and run ID, in the format entity/project/run_id.

property Run.project

Name of the W&B project associated with the run.

property Run.project_url

URL of the W&B project associated with the run, if there is one.

Offline runs do not have a project URL.

property Run.resumed

True if the run was resumed, False otherwise.

property Run.settings

A frozen copy of run’s Settings object.

property Run.start_time

Unix timestamp (in seconds) of when the run started.

property Run.sweep_id

Identifier for the sweep associated with the run, if there is one.

property Run.sweep_url

URL of the sweep associated with the run, if there is one.

Offline runs do not have a sweep URL.

property Run.tags

Tags associated with the run, if there are any.

property Run.url

The url for the W&B run, if there is one.

Offline runs will not have a url.

method Run.alert

alert(
    title: 'str',
    text: 'str',
    level: 'str | AlertLevel | None' = None,
    wait_duration: 'int | float | timedelta | None' = None
) → None

Create an alert with the given title and text.

Args:

  • title: The title of the alert, must be less than 64 characters long.
  • text: The text body of the alert.
  • level: The alert level to use, either: INFO, WARN, or ERROR.
  • wait_duration: The time to wait (in seconds) before sending another alert with this title.

method Run.define_metric

define_metric(
    name: 'str',
    step_metric: 'str | wandb_metric.Metric | None' = None,
    step_sync: 'bool | None' = None,
    hidden: 'bool | None' = None,
    summary: 'str | None' = None,
    goal: 'str | None' = None,
    overwrite: 'bool | None' = None
) → wandb_metric.Metric

Customize metrics logged with wandb.Run.log().

Args:

  • name: The name of the metric to customize.
  • step_metric: The name of another metric to serve as the X-axis for this metric in automatically generated charts.
  • step_sync: Automatically insert the last value of step_metric into wandb.Run.log() if it is not provided explicitly. Defaults to True if step_metric is specified.
  • hidden: Hide this metric from automatic plots.
  • summary: Specify aggregate metrics added to summary. Supported aggregations include “min”, “max”, “mean”, “last”, “best”, “copy” and “none”. “best” is used together with the goal parameter. “none” prevents a summary from being generated. “copy” is deprecated and should not be used.
  • goal: Specify how to interpret the “best” summary type. Supported options are “minimize” and “maximize”.
  • overwrite: If false, then this call is merged with previous define_metric calls for the same metric by using their values for any unspecified parameters. If true, then unspecified parameters overwrite values specified by previous calls.

Returns: An object that represents this call but can otherwise be discarded.

method Run.display

display(height: 'int' = 420, hidden: 'bool' = False) → bool

Display this run in Jupyter.

method Run.finish

finish(exit_code: 'int | None' = None, quiet: 'bool | None' = None) → None

Finish a run and upload any remaining data.

Marks the completion of a W&B run and ensures all data is synced to the server. The run’s final state is determined by its exit conditions and sync status.

Run States:

  • Running: Active run that is logging data and/or sending heartbeats.
  • Crashed: Run that stopped sending heartbeats unexpectedly.
  • Finished: Run completed successfully (exit_code=0) with all data synced.
  • Failed: Run completed with errors (exit_code!=0).
  • Killed: Run was forcibly stopped before it could finish.

Args:

  • exit_code: Integer indicating the run’s exit status. Use 0 for success, any other value marks the run as failed.
  • quiet: Deprecated. Configure logging verbosity using wandb.Settings(quiet=...).

method Run.finish_artifact

finish_artifact(
    artifact_or_path: 'Artifact | str',
    name: 'str | None' = None,
    type: 'str | None' = None,
    aliases: 'list[str] | None' = None,
    distributed_id: 'str | None' = None
) → Artifact

Finishes a non-finalized artifact as output of a run.

Subsequent “upserts” with the same distributed ID will result in a new version.

Args:

  • artifact_or_path: A path to the contents of this artifact, can be in the following forms: - /local/directory - /local/directory/file.txt - s3://bucket/path You can also pass an Artifact object created by calling wandb.Artifact.
  • name: An artifact name. May be prefixed with entity/project. Valid names can be in the following forms: - name:version - name:alias - digest This will default to the basename of the path prepended with the current run id if not specified.
  • type: The type of artifact to log, examples include dataset, model
  • aliases: Aliases to apply to this artifact, defaults to ["latest"]
  • distributed_id: Unique string that all distributed jobs share. If None, defaults to the run’s group name.

Returns: An Artifact object.

link_artifact(
    artifact: 'Artifact',
    target_path: 'str',
    aliases: 'list[str] | None' = None
) → Artifact | None

Link the given artifact to a portfolio (a promoted collection of artifacts).

Linked artifacts are visible in the UI for the specified portfolio.

Args:

  • artifact: the (public or local) artifact which will be linked
  • target_path: str - takes the following forms: {portfolio}, {project}/{portfolio}, or {entity}/{project}/{portfolio}
  • aliases: List[str] - optional alias(es) that will only be applied on this linked artifact inside the portfolio. The alias “latest” will always be applied to the latest version of an artifact that is linked.

Returns: The linked artifact if linking was successful, otherwise None.

link_model(
    path: 'StrPath',
    registered_model_name: 'str',
    name: 'str | None' = None,
    aliases: 'list[str] | None' = None
) → Artifact | None

Log a model artifact version and link it to a registered model in the model registry.

Linked model versions are visible in the UI for the specified registered model.

This method will:

  • Check if ’name’ model artifact has been logged. If so, use the artifact version that matches the files located at ‘path’ or log a new version. Otherwise log files under ‘path’ as a new model artifact, ’name’ of type ‘model’.
  • Check if registered model with name ‘registered_model_name’ exists in the ‘model-registry’ project. If not, create a new registered model with name ‘registered_model_name’.
  • Link version of model artifact ’name’ to registered model, ‘registered_model_name’.
  • Attach aliases from ‘aliases’ list to the newly linked model artifact version.

Args:

  • path: (str) A path to the contents of this model, can be in the following forms:
    • /local/directory
    • /local/directory/file.txt
    • s3://bucket/path
  • registered_model_name: The name of the registered model that the model is to be linked to. A registered model is a collection of model versions linked to the model registry, typically representing a team’s specific ML Task. The entity that this registered model belongs to will be derived from the run.
  • name: The name of the model artifact that files in ‘path’ will be logged to. This will default to the basename of the path prepended with the current run id if not specified.
  • aliases: Aliases that will only be applied on this linked artifact inside the registered model. The alias “latest” will always be applied to the latest version of an artifact that is linked.

Raises:

  • AssertionError: If registered_model_name is a path or if model artifact ’name’ is of a type that does not contain the substring ‘model’.
  • ValueError: If name has invalid special characters.

Returns: The linked artifact if linking was successful, otherwise None.

method Run.log

log(
    data: 'dict[str, Any]',
    step: 'int | None' = None,
    commit: 'bool | None' = None
) → None

Upload run data.

Use log to log data from runs, such as scalars, images, video, histograms, plots, and tables. See Log objects and media for code snippets, best practices, and more.

Basic usage:

import wandb

with wandb.init() as run:
     run.log({"train-loss": 0.5, "accuracy": 0.9})

The previous code snippet saves the loss and accuracy to the run’s history and updates the summary values for these metrics.

Visualize logged data in a workspace at wandb.ai, or locally on a self-hosted instance of the W&B app, or export data to visualize and explore locally, such as in a Jupyter notebook, with the Public API.

Logged values don’t have to be scalars. You can log any W&B supported Data Type such as images, audio, video, and more. For example, you can use wandb.Table to log structured data. See Log tables, visualize and query data tutorial for more details.

W&B organizes metrics with a forward slash (/) in their name into sections named using the text before the final slash. For example, the following results in two sections named “train” and “validate”:

with wandb.init() as run:
     # Log metrics in the "train" section.
     run.log(
         {
             "train/accuracy": 0.9,
             "train/loss": 30,
             "validate/accuracy": 0.8,
             "validate/loss": 20,
         }
     )

Only one level of nesting is supported; run.log({"a/b/c": 1}) produces a section named “a/b”.

run.log() is not intended to be called more than a few times per second. For optimal performance, limit your logging to once every N iterations, or collect data over multiple iterations and log it in a single step.

By default, each call to log creates a new “step”. The step must always increase, and it is not possible to log to a previous step. You can use any metric as the X axis in charts. See Custom log axes for more details.

In many cases, it is better to treat the W&B step like you’d treat a timestamp rather than a training step.

with wandb.init() as run:
     # Example: log an "epoch" metric for use as an X axis.
     run.log({"epoch": 40, "train-loss": 0.5})

It is possible to use multiple wandb.Run.log() invocations to log to the same step with the step and commit parameters. The following are all equivalent:

with wandb.init() as run:
     # Normal usage:
     run.log({"train-loss": 0.5, "accuracy": 0.8})
     run.log({"train-loss": 0.4, "accuracy": 0.9})

     # Implicit step without auto-incrementing:
     run.log({"train-loss": 0.5}, commit=False)
     run.log({"accuracy": 0.8})
     run.log({"train-loss": 0.4}, commit=False)
     run.log({"accuracy": 0.9})

     # Explicit step:
     run.log({"train-loss": 0.5}, step=current_step)
     run.log({"accuracy": 0.8}, step=current_step)
     current_step += 1
     run.log({"train-loss": 0.4}, step=current_step)
     run.log({"accuracy": 0.9}, step=current_step)

Args:

  • data: A dict with str keys and values that are serializable
  • Python objects including: int, float and string; any of the wandb.data_types; lists, tuples and NumPy arrays of serializable Python objects; other dicts of this structure.
  • step: The step number to log. If None, then an implicit auto-incrementing step is used. See the notes in the description.
  • commit: If true, finalize and upload the step. If false, then accumulate data for the step. See the notes in the description. If step is None, then the default is commit=True; otherwise, the default is commit=False.

Examples: For more and more detailed examples, see our guides to logging.

Basic usage

import wandb

with wandb.init() as run:
    run.log({"train-loss": 0.5, "accuracy": 0.9

Incremental logging

import wandb

with wandb.init() as run:
    run.log({"loss": 0.2}, commit=False)
    # Somewhere else when I'm ready to report this step:
    run.log({"accuracy": 0.8})

Histogram

import numpy as np
import wandb

# sample gradients at random from normal distribution
gradients = np.random.randn(100, 100)
with wandb.init() as run:
    run.log({"gradients": wandb.Histogram(gradients)})

Image from NumPy

import numpy as np
import wandb

with wandb.init() as run:
    examples = []
    for i in range(3):
         pixels = np.random.randint(low=0, high=256, size=(100, 100, 3))
         image = wandb.Image(pixels, caption=f"random field {i}")
         examples.append(image)
    run.log({"examples": examples})

Image from PIL

import numpy as np
from PIL import Image as PILImage
import wandb

with wandb.init() as run:
    examples = []
    for i in range(3):
         pixels = np.random.randint(
             low=0,
             high=256,
             size=(100, 100, 3),
             dtype=np.uint8,
         )
         pil_image = PILImage.fromarray(pixels, mode="RGB")
         image = wandb.Image(pil_image, caption=f"random field {i}")
         examples.append(image)
    run.log({"examples": examples})

Video from NumPy

import numpy as np
import wandb

with wandb.init() as run:
    # axes are (time, channel, height, width)
    frames = np.random.randint(
         low=0,
         high=256,
         size=(10, 3, 100, 100),
         dtype=np.uint8,
    )
    run.log({"video": wandb.Video(frames, fps=4)})

Matplotlib plot

from matplotlib import pyplot as plt
import numpy as np
import wandb

with wandb.init() as run:
    fig, ax = plt.subplots()
    x = np.linspace(0, 10)
    y = x * x
    ax.plot(x, y)  # plot y = x^2
    run.log({"chart": fig})

PR Curve

import wandb

with wandb.init() as run:
    run.log({"pr": wandb.plot.pr_curve(y_test, y_probas, labels)})

3D Object

import wandb

with wandb.init() as run:
    run.log(
         {
             "generated_samples": [
                 wandb.Object3D(open("sample.obj")),
                 wandb.Object3D(open("sample.gltf")),
                 wandb.Object3D(open("sample.glb")),
             ]
         }
    )

Raises:

  • wandb.Error: If called before wandb.init().
  • ValueError: If invalid data is passed.

method Run.log_artifact

log_artifact(
    artifact_or_path: 'Artifact | StrPath',
    name: 'str | None' = None,
    type: 'str | None' = None,
    aliases: 'list[str] | None' = None,
    tags: 'list[str] | None' = None
) → Artifact

Declare an artifact as an output of a run.

Args:

  • artifact_or_path: (str or Artifact) A path to the contents of this artifact, can be in the following forms: - /local/directory - /local/directory/file.txt - s3://bucket/path You can also pass an Artifact object created by calling wandb.Artifact.
  • name: (str, optional) An artifact name. Valid names can be in the following forms: - name:version - name:alias - digest This will default to the basename of the path prepended with the current run id if not specified.
  • type: (str) The type of artifact to log, examples include dataset, model
  • aliases: (list, optional) Aliases to apply to this artifact, defaults to ["latest"]
  • tags: (list, optional) Tags to apply to this artifact, if any.

Returns: An Artifact object.

method Run.log_code

log_code(
    root: 'str | None' = '.',
    name: 'str | None' = None,
    include_fn: 'Callable[[str, str], bool] | Callable[[str], bool]' = <function _is_py_requirements_or_dockerfile at 0x102da5f30>,
    exclude_fn: 'Callable[[str, str], bool] | Callable[[str], bool]' = <function exclude_wandb_fn at 0x103b4c5e0>
) → Artifact | None

Save the current state of your code to a W&B Artifact.

By default, it walks the current directory and logs all files that end with .py.

Args:

  • root: The relative (to os.getcwd()) or absolute path to recursively find code from.
  • name: (str, optional) The name of our code artifact. By default, we’ll name the artifact source-$PROJECT_ID-$ENTRYPOINT_RELPATH. There may be scenarios where you want many runs to share the same artifact. Specifying name allows you to achieve that.
  • include_fn: A callable that accepts a file path and (optionally) root path and returns True when it should be included and False otherwise. This
  • defaults to lambda path, root: path.endswith(".py").
  • exclude_fn: A callable that accepts a file path and (optionally) root path and returns True when it should be excluded and False otherwise. This defaults to a function that excludes all files within <root>/.wandb/ and <root>/wandb/ directories.

Examples: Basic usage

import wandb

with wandb.init() as run:
    run.log_code()

Advanced usage

import wandb

with wandb.init() as run:
    run.log_code(
         root="../",
         include_fn=lambda path: path.endswith(".py") or path.endswith(".ipynb"),
         exclude_fn=lambda path, root: os.path.relpath(path, root).startswith(
             "cache/"
         ),
    )

Returns: An Artifact object if code was logged

method Run.log_model

log_model(
    path: 'StrPath',
    name: 'str | None' = None,
    aliases: 'list[str] | None' = None
) → None

Logs a model artifact containing the contents inside the ‘path’ to a run and marks it as an output to this run.

The name of model artifact can only contain alphanumeric characters, underscores, and hyphens.

Args:

  • path: (str) A path to the contents of this model, can be in the following forms: - /local/directory - /local/directory/file.txt - s3://bucket/path
  • name: A name to assign to the model artifact that the file contents will be added to. This will default to the basename of the path prepended with the current run id if not specified.
  • aliases: Aliases to apply to the created model artifact, defaults to ["latest"]

Raises:

  • ValueError: If name has invalid special characters.

Returns: None

method Run.mark_preempting

mark_preempting() → None

Mark this run as preempting.

Also tells the internal process to immediately report this to server.

method Run.restore

restore(
    name: 'str',
    run_path: 'str | None' = None,
    replace: 'bool' = False,
    root: 'str | None' = None
) → None | TextIO

Download the specified file from cloud storage.

File is placed into the current directory or run directory. By default, will only download the file if it doesn’t already exist.

Args:

  • name: The name of the file.
  • run_path: Optional path to a run to pull files from, i.e. username/project_name/run_id if wandb.init has not been called, this is required.
  • replace: Whether to download the file even if it already exists locally
  • root: The directory to download the file to. Defaults to the current directory or the run directory if wandb.init was called.

Returns: None if it can’t find the file, otherwise a file object open for reading.

Raises:

  • CommError: If W&B can’t connect to the W&B backend.
  • ValueError: If the file is not found or can’t find run_path.

method Run.save

save(
    glob_str: 'str | os.PathLike',
    base_path: 'str | os.PathLike | None' = None,
    policy: 'PolicyName' = 'live'
) → bool | list[str]

Sync one or more files to W&B.

Relative paths are relative to the current working directory.

A Unix glob, such as “myfiles/*”, is expanded at the time save is called regardless of the policy. In particular, new files are not picked up automatically.

A base_path may be provided to control the directory structure of uploaded files. It should be a prefix of glob_str, and the directory structure beneath it is preserved.

When given an absolute path or glob and no base_path, one directory level is preserved as in the example above.

Args:

  • glob_str: A relative or absolute path or Unix glob.
  • base_path: A path to use to infer a directory structure; see examples.
  • policy: One of live, now, or end.
    • live: upload the file as it changes, overwriting the previous version
    • now: upload the file once now
    • end: upload file when the run ends

Returns: Paths to the symlinks created for the matched files.

For historical reasons, this may return a boolean in legacy code.

import wandb

run = wandb.init()

run.save("these/are/myfiles/*")
# => Saves files in a "these/are/myfiles/" folder in the run.

run.save("these/are/myfiles/*", base_path="these")
# => Saves files in an "are/myfiles/" folder in the run.

run.save("/User/username/Documents/run123/*.txt")
# => Saves files in a "run123/" folder in the run. See note below.

run.save("/User/username/Documents/run123/*.txt", base_path="/User")
# => Saves files in a "username/Documents/run123/" folder in the run.

run.save("files/*/saveme.txt")
# => Saves each "saveme.txt" file in an appropriate subdirectory
#    of "files/".

# Explicitly finish the run since a context manager is not used.
run.finish()

method Run.status

status() → RunStatus

Get sync info from the internal backend, about the current run’s sync status.

method Run.unwatch

unwatch(
    models: 'torch.nn.Module | Sequence[torch.nn.Module] | None' = None
) → None

Remove pytorch model topology, gradient and parameter hooks.

Args:

  • models: Optional list of pytorch models that have had watch called on them.

method Run.upsert_artifact

upsert_artifact(
    artifact_or_path: 'Artifact | str',
    name: 'str | None' = None,
    type: 'str | None' = None,
    aliases: 'list[str] | None' = None,
    distributed_id: 'str | None' = None
) → Artifact

Declare (or append to) a non-finalized artifact as output of a run.

Note that you must call run.finish_artifact() to finalize the artifact. This is useful when distributed jobs need to all contribute to the same artifact.

Args:

  • artifact_or_path: A path to the contents of this artifact, can be in the following forms:
    • /local/directory
    • /local/directory/file.txt
    • s3://bucket/path
  • name: An artifact name. May be prefixed with “entity/project”. Defaults to the basename of the path prepended with the current run ID if not specified. Valid names can be in the following forms:
    • name:version
    • name:alias
    • digest
  • type: The type of artifact to log. Common examples include dataset, model.
  • aliases: Aliases to apply to this artifact, defaults to ["latest"].
  • distributed_id: Unique string that all distributed jobs share. If None, defaults to the run’s group name.

Returns: An Artifact object.

method Run.use_artifact

use_artifact(
    artifact_or_name: 'str | Artifact',
    type: 'str | None' = None,
    aliases: 'list[str] | None' = None,
    use_as: 'str | None' = None
) → Artifact

Declare an artifact as an input to a run.

Call download or file on the returned object to get the contents locally.

Args:

  • artifact_or_name: The name of the artifact to use. May be prefixed with the name of the project the artifact was logged to ("" or “/”). 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: The type of artifact to use.
  • aliases: Aliases to apply to this artifact
  • use_as: This argument is deprecated and does nothing.

Returns: An Artifact object.

Examples:

import wandb

run = wandb.init(project="<example>")

# Use an artifact by name and alias
artifact_a = run.use_artifact(artifact_or_name="<name>:<alias>")

# Use an artifact by name and version
artifact_b = run.use_artifact(artifact_or_name="<name>:v<version>")

# Use an artifact by entity/project/name:alias
artifact_c = run.use_artifact(
   artifact_or_name="<entity>/<project>/<name>:<alias>"
)

# Use an artifact by entity/project/name:version
artifact_d = run.use_artifact(
   artifact_or_name="<entity>/<project>/<name>:v<version>"
)

# Explicitly finish the run since a context manager is not used.
run.finish()

method Run.use_model

use_model(name: 'str') → FilePathStr

Download the files logged in a model artifact ’name’.

Args:

  • name: A model artifact name. ’name’ must match the name of an existing logged model artifact. May be prefixed with entity/project/. Valid names can be in the following forms
    • model_artifact_name:version
    • model_artifact_name:alias

Returns:

  • path (str): Path to downloaded model artifact file(s).

Raises:

  • AssertionError: If model artifact ’name’ is of a type that does not contain the substring ‘model’.

method Run.watch

watch(
    models: 'torch.nn.Module | Sequence[torch.nn.Module]',
    criterion: 'torch.F | None' = None,
    log: "Literal['gradients', 'parameters', 'all'] | None" = 'gradients',
    log_freq: 'int' = 1000,
    idx: 'int | None' = None,
    log_graph: 'bool' = False
) → None

Hook into given PyTorch model to monitor gradients and the model’s computational graph.

This function can track parameters, gradients, or both during training.

Args:

  • models: A single model or a sequence of models to be monitored.
  • criterion: The loss function being optimized (optional).
  • log: Specifies whether to log “gradients”, “parameters”, or “all”. Set to None to disable logging. (default=“gradients”).
  • log_freq: Frequency (in batches) to log gradients and parameters. (default=1000)
  • idx: Index used when tracking multiple models with wandb.watch. (default=None)
  • log_graph: Whether to log the model’s computational graph. (default=False)

Raises: ValueError: If wandb.init() has not been called or if any of the models are not instances of torch.nn.Module.

2.3 - Settings

View the source code

Settings for the W&B SDK.

This class manages configuration settings for the W&B SDK, ensuring type safety and validation of all settings. Settings are accessible as attributes and can be initialized programmatically, through environment variables (WANDB_ prefix), and with configuration files.

The settings are organized into three categories:

  1. Public settings: Core configuration options that users can safely modify to customize W&B’s behavior for their specific needs.
  2. Internal settings: Settings prefixed with ‘x_’ that handle low-level SDK behavior. These settings are primarily for internal use and debugging. While they can be modified, they are not considered part of the public API and may change without notice in future versions.
  3. Computed settings: Read-only settings that are automatically derived from other settings or the environment.

Attributes:

  • allow_offline_artifacts (bool): Flag to allow table artifacts to be synced in offline mode. To revert to the old behavior, set this to False.

  • allow_val_change (bool): Flag to allow modification of Config values after they’ve been set.

  • anonymous (Optional): Controls anonymous data logging. Possible values are:

    • “never”: requires you to link your W&B account before tracking the run, so you don’t accidentally create an anonymous run.
    • “allow”: lets a logged-in user track runs with their account, but lets someone who is running the script without a W&B account see the charts in the UI.
    • “must”: sends the run to an anonymous account instead of to a signed-up user account.

  • api_key (Optional): The W&B API key.

  • azure_account_url_to_access_key (Optional): Mapping of Azure account URLs to their corresponding access keys for Azure integration.

  • base_url (str): The URL of the W&B backend for data synchronization.

  • code_dir (Optional): Directory containing the code to be tracked by W&B.

  • config_paths (Optional): Paths to files to load configuration from into the Config object.

  • console (Literal): The type of console capture to be applied. Possible values are: “auto” - Automatically selects the console capture method based on the system environment and settings. “off” - Disables console capture. “redirect” - Redirects low-level file descriptors for capturing output. “wrap” - Overrides the write methods of sys.stdout/sys.stderr. Will be mapped to either “wrap_raw” or “wrap_emu” based on the state of the system. “wrap_raw” - Same as “wrap” but captures raw output directly instead of through an emulator. Derived from the wrap setting and should not be set manually. “wrap_emu” - Same as “wrap” but captures output through an emulator. Derived from the wrap setting and should not be set manually.

  • console_multipart (bool): Whether to produce multipart console log files.

  • credentials_file (str): Path to file for writing temporary access tokens.

  • disable_code (bool): Whether to disable capturing the code.

  • disable_git (bool): Whether to disable capturing the git state.

  • disable_job_creation (bool): Whether to disable the creation of a job artifact for W&B Launch.

  • docker (Optional): The Docker image used to execute the script.

  • email (Optional): The email address of the user.

  • entity (Optional): The W&B entity, such as a user or a team.

  • force (bool): Whether to pass the force flag to wandb.login().

  • fork_from (Optional): Specifies a point in a previous execution of a run to fork from. The point is defined by the run ID, a metric, and its value. Currently, only the metric ‘_step’ is supported.

  • git_commit (Optional): The git commit hash to associate with the run.

  • git_remote (str): The git remote to associate with the run.

  • git_remote_url (Optional): The URL of the git remote repository.

  • git_root (Optional): Root directory of the git repository.

  • host (Optional): Hostname of the machine running the script.

  • http_proxy (Optional): Custom proxy servers for http requests to W&B.

  • https_proxy (Optional): Custom proxy servers for https requests to W&B.

  • identity_token_file (Optional): Path to file containing an identity token (JWT) for authentication.

  • ignore_globs (Sequence): Unix glob patterns relative to files_dir specifying files to exclude from upload.

  • init_timeout (float): Time in seconds to wait for the wandb.init call to complete before timing out.

  • insecure_disable_ssl (bool): Whether to insecurely disable SSL verification.

  • job_name (Optional): Name of the Launch job running the script.

  • job_source (Optional): Source type for Launch.

  • label_disable (bool): Whether to disable automatic labeling features.

  • launch_config_path (Optional): Path to the launch configuration file.

  • login_timeout (Optional): Time in seconds to wait for login operations before timing out.

  • mode (Literal): The operating mode for W&B logging and synchronization.

  • notebook_name (Optional): Name of the notebook if running in a Jupyter-like environment.

  • organization (Optional): The W&B organization.

  • program (Optional): Path to the script that created the run, if available.

  • program_abspath (Optional): The absolute path from the root repository directory to the script that created the run. Root repository directory is defined as the directory containing the .git directory, if it exists. Otherwise, it’s the current working directory.

  • program_relpath (Optional): The relative path to the script that created the run.

  • project (Optional): The W&B project ID.

  • quiet (bool): Flag to suppress non-essential output.

  • reinit (Union): What to do when wandb.init() is called while a run is active. Options:

    • “default”: Use “finish_previous” in notebooks and “return_previous” otherwise.
    • “return_previous”: Return the most recently created run that is not yet finished. This does not update wandb.run; see the “create_new” option.
    • “finish_previous”: Finish all active runs, then return a new run.
    • “create_new”: Create a new run without modifying other active runs. Does not update wandb.run and top-level functions like wandb.log. Because of this, some older integrations that rely on the global run will not work. Can also be a boolean, but this is deprecated. False is the same as “return_previous”, and True is the same as “finish_previous”.

  • relogin (bool): Flag to force a new login attempt.

  • resume (Optional): Specifies the resume behavior for the run. Options:

    • “must”: Resumes from an existing run with the same ID. If no such run exists, it will result in failure.
    • “allow”: Attempts to resume from an existing run with the same ID. If none is found, a new run will be created.
    • “never”: Always starts a new run. If a run with the same ID already exists, it will result in failure.
    • “auto”: Automatically resumes from the most recent failed run on the same machine.

  • resume_from (Optional): Specifies a point in a previous execution of a run to resume from. The point is defined by the run ID, a metric, and its value. Currently, only the metric ‘_step’ is supported.

  • root_dir (str): The root directory to use as the base for all run-related paths. In particular, this is used to derive the wandb directory and the run directory.

  • run_group (Optional): Group identifier for related runs. Used for grouping runs in the UI.

  • run_id (Optional): The ID of the run.

  • run_job_type (Optional): Type of job being run (e.g., training, evaluation).

  • run_name (Optional): Human-readable name for the run.

  • run_notes (Optional): Additional notes or description for the run.

  • run_tags (Optional): Tags to associate with the run for organization and filtering.

  • sagemaker_disable (bool): Flag to disable SageMaker-specific functionality.

  • save_code (Optional): Whether to save the code associated with the run.

  • settings_system (Optional): Path to the system-wide settings file.

  • show_errors (bool): Whether to display error messages.

  • show_info (bool): Whether to display informational messages.

  • show_warnings (bool): Whether to display warning messages.

  • silent (bool): Flag to suppress all output.

  • strict (Optional): Whether to enable strict mode for validation and error checking.

  • summary_timeout (int): Time in seconds to wait for summary operations before timing out.

  • sweep_id (Optional): Identifier of the sweep this run belongs to.

  • sweep_param_path (Optional): Path to the sweep parameters configuration.

  • symlink (bool): Whether to use symlinks (True by default except on Windows).

  • sync_tensorboard (Optional): Whether to synchronize TensorBoard logs with W&B.

  • table_raise_on_max_row_limit_exceeded (bool): Whether to raise an exception when table row limits are exceeded.

  • username (Optional): Username.

  • x_skip_transaction_log (bool): Whether to skip saving the run events to the transaction log. This is only relevant for online runs. Can be used to reduce the amount of data written to disk. Should be used with caution, as it removes the gurantees about recoverability.

  • x_stats_open_metrics_endpoints (Optional): OpenMetrics /metrics endpoints to monitor for system metrics.

  • x_stats_open_metrics_filters (Union): Filter to apply to metrics collected from OpenMetrics /metrics endpoints. Supports two formats:

    • {“metric regex pattern, including endpoint name as prefix”: {“label”: “label value regex pattern”}}
    • (“metric regex pattern 1”, “metric regex pattern 2”, …)

3 - Data Types

Defines Data Types for logging interactive visualizations to W&B.

3.1 - Audio

View the source code

class Audio

W&B class for audio clips.

method Audio.__init__

__init__(
    data_or_path: Union[str, pathlib.Path, list, ForwardRef('np.ndarray')],
    sample_rate: Optional[int] = None,
    caption: Optional[str] = None
)

Accept a path to an audio file or a numpy array of audio data.

Args:

  • data_or_path: A path to an audio file or a NumPy array of audio data.
  • sample_rate: Sample rate, required when passing in raw NumPy array of audio data.
  • caption: Caption to display with audio.

classmethod Audio.durations

durations(audio_list)

Calculate the duration of the audio files.

classmethod Audio.sample_rates

sample_rates(audio_list)

Get sample rates of the audio files.

3.2 - box3d()

View the source code

function box3d

box3d(
    center: 'npt.ArrayLike',
    size: 'npt.ArrayLike',
    orientation: 'npt.ArrayLike',
    color: 'RGBColor',
    label: 'Optional[str]' = None,
    score: 'Optional[numeric]' = None
) → Box3D

Returns a Box3D.

Args:

  • center: The center point of the box as a length-3 ndarray.
  • size: The box’s X, Y and Z dimensions as a length-3 ndarray.
  • orientation: The rotation transforming global XYZ coordinates into the box’s local XYZ coordinates, given as a length-4 ndarray [r, x, y, z] corresponding to the non-zero quaternion r + xi + yj + zk.
  • color: The box’s color as an (r, g, b) tuple with 0 <= r,g,b <= 1.
  • label: An optional label for the box.
  • score: An optional score for the box.

3.3 - Html

View the source code

class Html

W&B class for logging HTML content to W&B.

method Html.__init__

__init__(
    data: Union[str, pathlib.Path, ForwardRef('TextIO')],
    inject: bool = True,
    data_is_not_path: bool = False
) → None

Creates a W&B HTML object.

Args: data: A string that is a path to a file with the extension “.html”, or a string or IO object containing literal HTML.

  • inject: Add a stylesheet to the HTML object. If set to False the HTML will pass through unchanged.
  • data_is_not_path: If set to False, the data will be treated as a path to a file.

Examples: It can be initialized by providing a path to a file:

with wandb.init() as run:
    run.log({"html": wandb.Html("./index.html")})

Alternatively, it can be initialized by providing literal HTML, in either a string or IO object:

with wandb.init() as run:
    run.log({"html": wandb.Html("<h1>Hello, world!</h1>")})

3.4 - Image

View the source code

class Image

A class for logging images to W&B.

method Image.__init__

__init__(
    data_or_path: 'ImageDataOrPathType',
    mode: Optional[str] = None,
    caption: Optional[str] = None,
    grouping: Optional[int] = None,
    classes: Optional[ForwardRef('Classes'), Sequence[dict]] = None,
    boxes: Optional[Dict[str, ForwardRef('BoundingBoxes2D')], Dict[str, dict]] = None,
    masks: Optional[Dict[str, ForwardRef('ImageMask')], Dict[str, dict]] = None,
    file_type: Optional[str] = None,
    normalize: bool = True
) → None

Initialize a wandb.Image object.

Args:

  • data_or_path: Accepts NumPy array/pytorch tensor of image data, a PIL image object, or a path to an image file. If a NumPy array or pytorch tensor is provided, the image data will be saved to the given file type. If the values are not in the range [0, 255] or all values are in the range [0, 1], the image pixel values will be normalized to the range [0, 255] unless normalize is set to False.
    • pytorch tensor should be in the format (channel, height, width)
    • NumPy array should be in the format (height, width, channel)
  • mode: The PIL mode for an image. Most common are “L”, “RGB”,
  • "RGBA". Full explanation at https: //pillow.readthedocs.io/en/stable/handbook/concepts.html#modes
  • caption: Label for display of image.
  • grouping: The grouping number for the image.
  • classes: A list of class information for the image, used for labeling bounding boxes, and image masks.
  • boxes: A dictionary containing bounding box information for the image.
  • see: https://docs.wandb.ai/ref/python/data-types/boundingboxes2d/
  • masks: A dictionary containing mask information for the image.
  • see: https://docs.wandb.ai/ref/python/data-types/imagemask/
  • file_type: The file type to save the image as. This parameter has no effect if data_or_path is a path to an image file.
  • normalize: If True, normalize the image pixel values to fall within the range of [0, 255]. Normalize is only applied if data_or_path is a numpy array or pytorch tensor.

Examples: Create a wandb.Image from a numpy array

import numpy as np
import wandb

with wandb.init() as run:
    examples = []
    for i in range(3):
         pixels = np.random.randint(low=0, high=256, size=(100, 100, 3))
         image = wandb.Image(pixels, caption=f"random field {i}")
         examples.append(image)
    run.log({"examples": examples})

Create a wandb.Image from a PILImage

import numpy as np
from PIL import Image as PILImage
import wandb

with wandb.init() as run:
    examples = []
    for i in range(3):
         pixels = np.random.randint(
             low=0, high=256, size=(100, 100, 3), dtype=np.uint8
         )
         pil_image = PILImage.fromarray(pixels, mode="RGB")
         image = wandb.Image(pil_image, caption=f"random field {i}")
         examples.append(image)
    run.log({"examples": examples})

Log .jpg rather than .png (default)

import numpy as np
import wandb

with wandb.init() as run:
    examples = []
    for i in range(3):
         pixels = np.random.randint(low=0, high=256, size=(100, 100, 3))
         image = wandb.Image(
             pixels, caption=f"random field {i}", file_type="jpg"
         )
         examples.append(image)
    run.log({"examples": examples})

property Image.image

3.5 - Molecule

View the source code

class Molecule

W&B class for 3D Molecular data.

method Molecule.__init__

__init__(
    data_or_path: Union[str, pathlib.Path, ForwardRef('TextIO')],
    caption: Optional[str] = None,
    **kwargs: str
) → None

Initialize a Molecule object.

Args:

  • data_or_path: Molecule can be initialized from a file name or an io object.
  • caption: Caption associated with the molecule for display.

3.6 - Object3D

View the source code

class Object3D

W&B class for 3D point clouds.

method Object3D.__init__

__init__(
    data_or_path: Union[ForwardRef('np.ndarray'), str, pathlib.Path, ForwardRef('TextIO'), dict],
    caption: Optional[str] = None,
    **kwargs: Optional[str, ForwardRef('FileFormat3D')]
) → None

Creates a W&B Object3D object.

Args:

  • data_or_path: Object3D can be initialized from a file or a numpy array.
  • caption: Caption associated with the object for display.

Examples: The shape of the numpy array must be one of either

[[x y z],       ...] nx3
[[x y z c],     ...] nx4 where c is a category with supported range [1, 14]
[[x y z r g b], ...] nx6 where is rgb is color

3.7 - Plotly

View the source code

class Plotly

W&B class for Plotly plots.

method Plotly.__init__

__init__(
    val: Union[ForwardRef('plotly.Figure'), ForwardRef('matplotlib.artist.Artist')]
)

Initialize a Plotly object.

Args:

  • val: Matplotlib or Plotly figure.

3.8 - Table

View the source code

class Table

The Table class used to display and analyze tabular data.

Unlike traditional spreadsheets, Tables support numerous types of data: scalar values, strings, numpy arrays, and most subclasses of wandb.data_types.Media. This means you can embed Images, Video, Audio, and other sorts of rich, annotated media directly in Tables, alongside other traditional scalar values.

This class is the primary class used to generate W&B Tables https://docs.wandb.ai/guides/models/tables/.

method Table.__init__

__init__(
    columns=None,
    data=None,
    rows=None,
    dataframe=None,
    dtype=None,
    optional=True,
    allow_mixed_types=False,
    log_mode: Optional[Literal['IMMUTABLE', 'MUTABLE', 'INCREMENTAL']] = 'IMMUTABLE'
)

Initializes a Table object.

The rows is available for legacy reasons and should not be used. The Table class uses data to mimic the Pandas API.

Args:

  • columns: (List[str]) Names of the columns in the table. Defaults to [“Input”, “Output”, “Expected”].
  • data: (List[List[any]]) 2D row-oriented array of values.
  • dataframe: (pandas.DataFrame) DataFrame object used to create the table. When set, data and columns arguments are ignored.
  • rows: (List[List[any]]) 2D row-oriented array of values.
  • optional: (Union[bool,List[bool]]) Determines if None values are allowed. Default to True - If a singular bool value, then the optionality is enforced for all columns specified at construction time - If a list of bool values, then the optionality is applied to each column - should be the same length as columns applies to all columns. A list of bool values applies to each respective column.
  • allow_mixed_types: (bool) Determines if columns are allowed to have mixed types (disables type validation). Defaults to False
  • log_mode: Optional[str] Controls how the Table is logged when mutations occur. Options: - “IMMUTABLE” (default): Table can only be logged once; subsequent logging attempts after the table has been mutated will be no-ops. - “MUTABLE”: Table can be re-logged after mutations, creating a new artifact version each time it’s logged. - “INCREMENTAL”: Table data is logged incrementally, with each log creating a new artifact entry containing the new data since the last log.

method Table.add_column

add_column(name, data, optional=False)

Adds a column of data to the table.

Args:

  • name: (str) - the unique name of the column
  • data: (list | np.array) - a column of homogeneous data
  • optional: (bool) - if null-like values are permitted

method Table.add_computed_columns

add_computed_columns(fn)

Adds one or more computed columns based on existing data.

Args:

  • fn: A function which accepts one or two parameters, ndx (int) and row (dict), which is expected to return a dict representing new columns for that row, keyed by the new column names.
    • ndx is an integer representing the index of the row. Only included if include_ndx is set to True.
    • row is a dictionary keyed by existing columns

method Table.add_data

add_data(*data)

Adds a new row of data to the table.

The maximum amount ofrows in a table is determined by wandb.Table.MAX_ARTIFACT_ROWS.

The length of the data should match the length of the table column.

method Table.add_row

add_row(*row)

Deprecated. Use Table.add_data method instead.

method Table.cast

cast(col_name, dtype, optional=False)

Casts a column to a specific data type.

This can be one of the normal python classes, an internal W&B type, or an example object, like an instance of wandb.Image or wandb.Classes.

Args:

  • col_name (str): The name of the column to cast.
  • dtype (class, wandb.wandb_sdk.interface._dtypes.Type, any): The target dtype.
  • optional (bool): If the column should allow Nones.

method Table.get_column

get_column(name, convert_to=None)

Retrieves a column from the table and optionally converts it to a NumPy object.

Args:

  • name: (str) - the name of the column
  • convert_to: (str, optional) - “numpy”: will convert the underlying data to numpy object

method Table.get_dataframe

get_dataframe()

Returns a pandas.DataFrame of the table.

method Table.get_index

get_index()

Returns an array of row indexes for use in other tables to create links.

3.9 - Video

View the source code

class Video

A class for logging videos to W&B.

method Video.__init__

__init__(
    data_or_path: Union[str, pathlib.Path, ForwardRef('np.ndarray'), ForwardRef('TextIO'), ForwardRef('BytesIO')],
    caption: Optional[str] = None,
    fps: Optional[int] = None,
    format: Optional[Literal['gif', 'mp4', 'webm', 'ogg']] = None
)

Initialize a W&B Video object.

Args:

  • data_or_path: Video can be initialized with a path to a file or an io object. Video can be initialized with a numpy tensor. The numpy tensor must be either 4 dimensional or 5 dimensional. The dimensions should be (number of frames, channel, height, width) or (batch, number of frames, channel, height, width) The format parameter must be specified with the format argument when initializing with a numpy array or io object.
  • caption: Caption associated with the video for display.
  • fps: The frame rate to use when encoding raw video frames. Default value is 4. This parameter has no effect when data_or_path is a string, or bytes.
  • format: Format of video, necessary if initializing with a numpy array or io object. This parameter will be used to determine the format to use when encoding the video data. Accepted values are “gif”, “mp4”, “webm”, or “ogg”. If no value is provided, the default format will be “gif”.

Examples: Log a numpy array as a video

import numpy as np
import wandb

with wandb.init() as run:
    # axes are (number of frames, channel, height, width)
    frames = np.random.randint(
         low=0, high=256, size=(10, 3, 100, 100), dtype=np.uint8
    )
    run.log({"video": wandb.Video(frames, format="mp4", fps=4)})

4 - Custom Charts

Create custom charts and visualizations.

4.1 - bar()

View the source code

function bar

bar(
    table: 'wandb.Table',
    label: 'str',
    value: 'str',
    title: 'str' = '',
    split_table: 'bool' = False
) → CustomChart

Constructs a bar chart from a wandb.Table of data.

Args:

  • table: A table containing the data for the bar chart.
  • label: The name of the column to use for the labels of each bar.
  • value: The name of the column to use for the values of each bar.
  • title: The title of the bar chart.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Example:

import random
import wandb

# Generate random data for the table
data = [
    ["car", random.uniform(0, 1)],
    ["bus", random.uniform(0, 1)],
    ["road", random.uniform(0, 1)],
    ["person", random.uniform(0, 1)],
]

# Create a table with the data
table = wandb.Table(data=data, columns=["class", "accuracy"])

# Initialize a W&B run and log the bar plot
with wandb.init(project="bar_chart") as run:
    # Create a bar plot from the table
    bar_plot = wandb.plot.bar(
         table=table,
         label="class",
         value="accuracy",
         title="Object Classification Accuracy",
    )

    # Log the bar chart to W&B
    run.log({"bar_plot": bar_plot})

4.2 - confusion_matrix()

View the source code

function confusion_matrix

confusion_matrix(
    probs: 'Sequence[Sequence[float]] | None' = None,
    y_true: 'Sequence[T] | None' = None,
    preds: 'Sequence[T] | None' = None,
    class_names: 'Sequence[str] | None' = None,
    title: 'str' = 'Confusion Matrix Curve',
    split_table: 'bool' = False
) → CustomChart

Constructs a confusion matrix from a sequence of probabilities or predictions.

Args:

  • probs: A sequence of predicted probabilities for each class. The sequence shape should be (N, K) where N is the number of samples and K is the number of classes. If provided, preds should not be provided.
  • y_true: A sequence of true labels.
  • preds: A sequence of predicted class labels. If provided, probs should not be provided.
  • class_names: Sequence of class names. If not provided, class names will be defined as “Class_1”, “Class_2”, etc.
  • title: Title of the confusion matrix chart.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Raises:

  • ValueError: If both probs and preds are provided or if the number of predictions and true labels are not equal. If the number of unique predicted classes exceeds the number of class names or if the number of unique true labels exceeds the number of class names.
  • wandb.Error: If numpy is not installed.

Examples: Logging a confusion matrix with random probabilities for wildlife classification:

import numpy as np
import wandb

# Define class names for wildlife
wildlife_class_names = ["Lion", "Tiger", "Elephant", "Zebra"]

# Generate random true labels (0 to 3 for 10 samples)
wildlife_y_true = np.random.randint(0, 4, size=10)

# Generate random probabilities for each class (10 samples x 4 classes)
wildlife_probs = np.random.rand(10, 4)
wildlife_probs = np.exp(wildlife_probs) / np.sum(
    np.exp(wildlife_probs),
    axis=1,
    keepdims=True,
)

# Initialize W&B run and log confusion matrix
with wandb.init(project="wildlife_classification") as run:
    confusion_matrix = wandb.plot.confusion_matrix(
         probs=wildlife_probs,
         y_true=wildlife_y_true,
         class_names=wildlife_class_names,
         title="Wildlife Classification Confusion Matrix",
    )
    run.log({"wildlife_confusion_matrix": confusion_matrix})

In this example, random probabilities are used to generate a confusion matrix.

Logging a confusion matrix with simulated model predictions and 85% accuracy:

import numpy as np
import wandb

# Define class names for wildlife
wildlife_class_names = ["Lion", "Tiger", "Elephant", "Zebra"]

# Simulate true labels for 200 animal images (imbalanced distribution)
wildlife_y_true = np.random.choice(
    [0, 1, 2, 3],
    size=200,
    p=[0.2, 0.3, 0.25, 0.25],
)

# Simulate model predictions with 85% accuracy
wildlife_preds = [
    y_t
    if np.random.rand() < 0.85
    else np.random.choice([x for x in range(4) if x != y_t])
    for y_t in wildlife_y_true
]

# Initialize W&B run and log confusion matrix
with wandb.init(project="wildlife_classification") as run:
    confusion_matrix = wandb.plot.confusion_matrix(
         preds=wildlife_preds,
         y_true=wildlife_y_true,
         class_names=wildlife_class_names,
         title="Simulated Wildlife Classification Confusion Matrix",
    )
    run.log({"wildlife_confusion_matrix": confusion_matrix})

In this example, predictions are simulated with 85% accuracy to generate a confusion matrix.

4.3 - histogram()

View the source code

function histogram

histogram(
    table: 'wandb.Table',
    value: 'str',
    title: 'str' = '',
    split_table: 'bool' = False
) → CustomChart

Constructs a histogram chart from a W&B Table.

Args:

  • table: The W&B Table containing the data for the histogram.
  • value: The label for the bin axis (x-axis).
  • title: The title of the histogram plot.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Example:

import math
import random
import wandb

# Generate random data
data = [[i, random.random() + math.sin(i / 10)] for i in range(100)]

# Create a W&B Table
table = wandb.Table(
    data=data,
    columns=["step", "height"],
)

# Create a histogram plot
histogram = wandb.plot.histogram(
    table,
    value="height",
    title="My Histogram",
)

# Log the histogram plot to W&B
with wandb.init(...) as run:
    run.log({"histogram-plot1": histogram})

4.4 - line_series()

View the source code

function line_series

line_series(
    xs: 'Iterable[Iterable[Any]] | Iterable[Any]',
    ys: 'Iterable[Iterable[Any]]',
    keys: 'Iterable[str] | None' = None,
    title: 'str' = '',
    xname: 'str' = 'x',
    split_table: 'bool' = False
) → CustomChart

Constructs a line series chart.

Args:

  • xs: Sequence of x values. If a singular array is provided, all y values are plotted against that x array. If an array of arrays is provided, each y value is plotted against the corresponding x array.
  • ys: Sequence of y values, where each iterable represents a separate line series.
  • keys: Sequence of keys for labeling each line series. If not provided, keys will be automatically generated as “line_1”, “line_2”, etc.
  • title: Title of the chart.
  • xname: Label for the x-axis.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Examples: Logging a single x array where all y series are plotted against the same x values:

import wandb

# Initialize W&B run
with wandb.init(project="line_series_example") as run:
    # x values shared across all y series
    xs = list(range(10))

    # Multiple y series to plot
    ys = [
         [i for i in range(10)],  # y = x
         [i**2 for i in range(10)],  # y = x^2
         [i**3 for i in range(10)],  # y = x^3
    ]

    # Generate and log the line series chart
    line_series_chart = wandb.plot.line_series(
         xs,
         ys,
         title="title",
         xname="step",
    )
    run.log({"line-series-single-x": line_series_chart})

In this example, a single xs series (shared x-values) is used for all ys series. This results in each y-series being plotted against the same x-values (0-9).

Logging multiple x arrays where each y series is plotted against its corresponding x array:

import wandb

# Initialize W&B run
with wandb.init(project="line_series_example") as run:
    # Separate x values for each y series
    xs = [
         [i for i in range(10)],  # x for first series
         [2 * i for i in range(10)],  # x for second series (stretched)
         [3 * i for i in range(10)],  # x for third series (stretched more)
    ]

    # Corresponding y series
    ys = [
         [i for i in range(10)],  # y = x
         [i**2 for i in range(10)],  # y = x^2
         [i**3 for i in range(10)],  # y = x^3
    ]

    # Generate and log the line series chart
    line_series_chart = wandb.plot.line_series(
         xs, ys, title="Multiple X Arrays Example", xname="Step"
    )
    run.log({"line-series-multiple-x": line_series_chart})

In this example, each y series is plotted against its own unique x series. This allows for more flexibility when the x values are not uniform across the data series.

Customizing line labels using keys:

import wandb

# Initialize W&B run
with wandb.init(project="line_series_example") as run:
    xs = list(range(10))  # Single x array
    ys = [
         [i for i in range(10)],  # y = x
         [i**2 for i in range(10)],  # y = x^2
         [i**3 for i in range(10)],  # y = x^3
    ]

    # Custom labels for each line
    keys = ["Linear", "Quadratic", "Cubic"]

    # Generate and log the line series chart
    line_series_chart = wandb.plot.line_series(
         xs,
         ys,
         keys=keys,  # Custom keys (line labels)
         title="Custom Line Labels Example",
         xname="Step",
    )
    run.log({"line-series-custom-keys": line_series_chart})

This example shows how to provide custom labels for the lines using the keys argument. The keys will appear in the legend as “Linear”, “Quadratic”, and “Cubic”.

4.5 - line()

View the source code

function line

line(
    table: 'wandb.Table',
    x: 'str',
    y: 'str',
    stroke: 'str | None' = None,
    title: 'str' = '',
    split_table: 'bool' = False
) → CustomChart

Constructs a customizable line chart.

Args:

  • table: The table containing data for the chart.
  • x: Column name for the x-axis values.
  • y: Column name for the y-axis values.
  • stroke: Column name to differentiate line strokes (e.g., for grouping lines).
  • title: Title of the chart.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Example:

import math
import random
import wandb

# Create multiple series of data with different patterns
data = []
for i in range(100):
     # Series 1: Sinusoidal pattern with random noise
     data.append([i, math.sin(i / 10) + random.uniform(-0.1, 0.1), "series_1"])
     # Series 2: Cosine pattern with random noise
     data.append([i, math.cos(i / 10) + random.uniform(-0.1, 0.1), "series_2"])
     # Series 3: Linear increase with random noise
     data.append([i, i / 10 + random.uniform(-0.5, 0.5), "series_3"])

# Define the columns for the table
table = wandb.Table(data=data, columns=["step", "value", "series"])

# Initialize wandb run and log the line chart
with wandb.init(project="line_chart_example") as run:
     line_chart = wandb.plot.line(
         table=table,
         x="step",
         y="value",
         stroke="series",  # Group by the "series" column
         title="Multi-Series Line Plot",
     )
     run.log({"line-chart": line_chart})

4.6 - plot_table()

View the source code

function plot_table

plot_table(
    vega_spec_name: 'str',
    data_table: 'wandb.Table',
    fields: 'dict[str, Any]',
    string_fields: 'dict[str, Any] | None' = None,
    split_table: 'bool' = False
) → CustomChart

Creates a custom charts using a Vega-Lite specification and a wandb.Table.

This function creates a custom chart based on a Vega-Lite specification and a data table represented by a wandb.Table object. The specification needs to be predefined and stored in the W&B backend. The function returns a custom chart object that can be logged to W&B using wandb.Run.log().

Args:

  • vega_spec_name: The name or identifier of the Vega-Lite spec that defines the visualization structure.
  • data_table: A wandb.Table object containing the data to be visualized.
  • fields: A mapping between the fields in the Vega-Lite spec and the corresponding columns in the data table to be visualized.
  • string_fields: A dictionary for providing values for any string constants required by the custom visualization.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass the chart object as argument to wandb.Run.log().

Raises:

  • wandb.Error: If data_table is not a wandb.Table object.

Example:

# Create a custom chart using a Vega-Lite spec and the data table.
import wandb

data = [[1, 1], [2, 2], [3, 3], [4, 4], [5, 5]]
table = wandb.Table(data=data, columns=["x", "y"])
fields = {"x": "x", "y": "y", "title": "MY TITLE"}

with wandb.init() as run:
   # Training code goes here

   # Create a custom title with `string_fields`.
   my_custom_chart = wandb.plot_table(
        vega_spec_name="wandb/line/v0",
        data_table=table,
        fields=fields,
        string_fields={"title": "Title"},
   )

   run.log({"custom_chart": my_custom_chart})

4.7 - pr_curve()

View the source code

function pr_curve

pr_curve(
    y_true: 'Iterable[T] | None' = None,
    y_probas: 'Iterable[numbers.Number] | None' = None,
    labels: 'list[str] | None' = None,
    classes_to_plot: 'list[T] | None' = None,
    interp_size: 'int' = 21,
    title: 'str' = 'Precision-Recall Curve',
    split_table: 'bool' = False
) → CustomChart

Constructs a Precision-Recall (PR) curve.

The Precision-Recall curve is particularly useful for evaluating classifiers on imbalanced datasets. A high area under the PR curve signifies both high precision (a low false positive rate) and high recall (a low false negative rate). The curve provides insights into the balance between false positives and false negatives at various threshold levels, aiding in the assessment of a model’s performance.

Args:

  • y_true: True binary labels. The shape should be (num_samples,).
  • y_probas: Predicted scores or probabilities for each class. These can be probability estimates, confidence scores, or non-thresholded decision values. The shape should be (num_samples, num_classes).
  • labels: Optional list of class names to replace numeric values in y_true for easier plot interpretation. For example, labels = ['dog', 'cat', 'owl'] will replace 0 with ‘dog’, 1 with ‘cat’, and 2 with ‘owl’ in the plot. If not provided, numeric values from y_true will be used.
  • classes_to_plot: Optional list of unique class values from y_true to be included in the plot. If not specified, all unique classes in y_true will be plotted.
  • interp_size: Number of points to interpolate recall values. The recall values will be fixed to interp_size uniformly distributed points in the range [0, 1], and the precision will be interpolated accordingly.
  • title: Title of the plot. Defaults to “Precision-Recall Curve”.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Raises:

  • wandb.Error: If NumPy, pandas, or scikit-learn is not installed.

Example:

import wandb

# Example for spam detection (binary classification)
y_true = [0, 1, 1, 0, 1]  # 0 = not spam, 1 = spam
y_probas = [
    [0.9, 0.1],  # Predicted probabilities for the first sample (not spam)
    [0.2, 0.8],  # Second sample (spam), and so on
    [0.1, 0.9],
    [0.8, 0.2],
    [0.3, 0.7],
]

labels = ["not spam", "spam"]  # Optional class names for readability

with wandb.init(project="spam-detection") as run:
    pr_curve = wandb.plot.pr_curve(
         y_true=y_true,
         y_probas=y_probas,
         labels=labels,
         title="Precision-Recall Curve for Spam Detection",
    )
    run.log({"pr-curve": pr_curve})

4.8 - roc_curve()

View the source code

function roc_curve

roc_curve(
    y_true: 'Sequence[numbers.Number]',
    y_probas: 'Sequence[Sequence[float]] | None' = None,
    labels: 'list[str] | None' = None,
    classes_to_plot: 'list[numbers.Number] | None' = None,
    title: 'str' = 'ROC Curve',
    split_table: 'bool' = False
) → CustomChart

Constructs Receiver Operating Characteristic (ROC) curve chart.

Args:

  • y_true: The true class labels (ground truth) for the target variable. Shape should be (num_samples,).
  • y_probas: The predicted probabilities or decision scores for each class. Shape should be (num_samples, num_classes).
  • labels: Human-readable labels corresponding to the class indices in y_true. For example, if labels=['dog', 'cat'], class 0 will be displayed as ‘dog’ and class 1 as ‘cat’ in the plot. If None, the raw class indices from y_true will be used. Default is None.
  • classes_to_plot: A subset of unique class labels to include in the ROC curve. If None, all classes in y_true will be plotted. Default is None.
  • title: Title of the ROC curve plot. Default is “ROC Curve”.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Raises:

  • wandb.Error: If numpy, pandas, or scikit-learn are not found.

Example:

import numpy as np
import wandb

# Simulate a medical diagnosis classification problem with three diseases
n_samples = 200
n_classes = 3

# True labels: assign "Diabetes", "Hypertension", or "Heart Disease" to
# each sample
disease_labels = ["Diabetes", "Hypertension", "Heart Disease"]
# 0: Diabetes, 1: Hypertension, 2: Heart Disease
y_true = np.random.choice([0, 1, 2], size=n_samples)

# Predicted probabilities: simulate predictions, ensuring they sum to 1
# for each sample
y_probas = np.random.dirichlet(np.ones(n_classes), size=n_samples)

# Specify classes to plot (plotting all three diseases)
classes_to_plot = [0, 1, 2]

# Initialize a W&B run and log a ROC curve plot for disease classification
with wandb.init(project="medical_diagnosis") as run:
   roc_plot = wandb.plot.roc_curve(
        y_true=y_true,
        y_probas=y_probas,
        labels=disease_labels,
        classes_to_plot=classes_to_plot,
        title="ROC Curve for Disease Classification",
   )
   run.log({"roc-curve": roc_plot})

4.9 - scatter()

View the source code

function scatter

scatter(
    table: 'wandb.Table',
    x: 'str',
    y: 'str',
    title: 'str' = '',
    split_table: 'bool' = False
) → CustomChart

Constructs a scatter plot from a wandb.Table of data.

Args:

  • table: The W&B Table containing the data to visualize.
  • x: The name of the column used for the x-axis.
  • y: The name of the column used for the y-axis.
  • title: The title of the scatter chart.
  • split_table: Whether the table should be split into a separate section in the W&B UI. If True, the table will be displayed in a section named “Custom Chart Tables”. Default is False.

Returns:

  • CustomChart: A custom chart object that can be logged to W&B. To log the chart, pass it to wandb.log().

Example:

import math
import random
import wandb

# Simulate temperature variations at different altitudes over time
data = [
   [i, random.uniform(-10, 20) - 0.005 * i + 5 * math.sin(i / 50)]
   for i in range(300)
]

# Create W&B table with altitude (m) and temperature (°C) columns
table = wandb.Table(data=data, columns=["altitude (m)", "temperature (°C)"])

# Initialize W&B run and log the scatter plot
with wandb.init(project="temperature-altitude-scatter") as run:
   # Create and log the scatter plot
   scatter_plot = wandb.plot.scatter(
        table=table,
        x="altitude (m)",
        y="temperature (°C)",
        title="Altitude vs Temperature",
   )
   run.log({"altitude-temperature-scatter": scatter_plot})