Skip to content

Utilities

autogenerate_seed()

Automatically generate a random seed for machine-learning experiments.

Source code in wandb_addons/utils.py 
77
78
79
80
81
82
83
def autogenerate_seed():
    """Automatically generate a random seed for machine-learning experiments."""
    max_seed = int(1024 * 1024 * 1024)
    seed = random.randint(1, max_seed)
    seed = -seed if seed < 0 else seed
    seed = seed % max_seed
    return seed

fetch_wandb_artifact(artifact_address, artifact_type)

Utility function for fetching a Weights & Biases artifact irrespective of whether a run has been initialized or not.

Parameters:

Name Type Description Default
artifact_address str

A human-readable name for the artifact, which is how you can identify the artifact in the UI or reference it in use_artifact calls. Names can contain letters, numbers, underscores, hyphens, and dots. The name must be unique across a project.

 required
artifact_type str

The type of the artifact, which is used to organize and differentiate artifacts. Common typesCinclude dataset or model, but you can use any string containing letters, numbers, underscores, hyphens, and dots.

 required

Returns:

Type Description
FilePathStr

The path to the downloaded contents.
Source code in wandb_addons/utils.py 
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
def fetch_wandb_artifact(artifact_address: str, artifact_type: str) -> FilePathStr:
    """Utility function for fetching a
    [Weights & Biases artifact](https://docs.wandb.ai/guides/artifacts)
    irrespective of whether a [run](https://docs.wandb.ai/guides/runs) has been initialized or not.

    Args:
        artifact_address (str): A human-readable name for the artifact, which is how you can
            identify the artifact in the UI or reference it in
            [`use_artifact`](https://docs.wandb.ai/ref/python/run#use_artifact) calls. Names can
            contain letters, numbers, underscores, hyphens, and dots. The name must be unique across
            a project.
        artifact_type (str): The type of the artifact, which is used to organize and differentiate
            artifacts. Common typesCinclude dataset or model, but you can use any string containing
            letters, numbers, underscores, hyphens, and dots.

    Returns:
        (wandb.util.FilePathStr): The path to the downloaded contents.
    """
    return (
        wandb.Api().artifact(artifact_address, type=artifact_type).download()
        if wandb.run is None
        else wandb.use_artifact(artifact_address, type=artifact_type).download()
    )

flatten_nested_dictionaries(d, parent_key='', sep='/')

A recursive function for flattening nested dictionaries.

Reference:

Answer to
[**Flatten nested dictionaries, compressing keys**](https://stackoverflow.com/q/6027558)
on StackOverflow: [stackoverflow.com/a/6027615](https://stackoverflow.com/a/6027615)

Parameters:

Name Type Description Default
d Dict

The input nested dictionary.

 required
parent_key str

The parent key.

 ''
sep str

The separator to use for the flattened keys.

 '/'

Returns:

Type Description
Dict

The flattened dictionary.
Source code in wandb_addons/utils.py 
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
def flatten_nested_dictionaries(d: Dict, parent_key: str = "", sep: str = "/") -> Dict:
    """A recursive function for flattening nested dictionaries.

    # Reference:
        Answer to
        [**Flatten nested dictionaries, compressing keys**](https://stackoverflow.com/q/6027558)
        on StackOverflow: [stackoverflow.com/a/6027615](https://stackoverflow.com/a/6027615)

    Args:
        d (Dict): The input nested dictionary.
        parent_key (str): The parent key.
        sep (str): The separator to use for the flattened keys.

    Returns:
        (Dict): The flattened dictionary.
    """
    items = []
    for k, v in d.items():
        new_key = parent_key + sep + k if parent_key else k
        if isinstance(v, MutableMapping):
            items.extend(flatten_nested_dictionaries(v, new_key, sep=sep).items())
        else:
            items.append((new_key, v))
    return dict(items)