log_configs()

Python package: neptune-scale

Logs the specified metadata to a Neptune run.

You can log configurations or other single values. Pass the data as a dataclass or a mapping-like object, where:

• keys are strings specifying the path to the metadata in the run structure.
• values are the configurations or other single values to log.

For example, {"parameters/learning_rate": 0.001}. In the attribute path, each forward slash / nests the attribute under a namespace. Use namespaces to structure the metadata into meaningful categories.

Related

• Log parameters and model configuration
• Namespaces and attributes

Parameters

data

Dict[str, float | bool | int | str | datetime | list | set | tuple]

optional

default: None

Configs or other values to log, as a dict, dataclass, or other mapping-like object with an .items() method.

Available types: float, integer, Boolean, string, and datetime.

Any datetime values that don't have the tzinfo attribute set are assumed to be in the local timezone.

flatten

bool

optional

default: False

Flatten nested dictionaries and dataclasses before logging.

cast_unsupported

bool

optional

default: False

Cast unsupported types to strings before logging.

Examples

Create a run and log metadata:

from neptune_scale import Run


if __name__ == "__main__":
    run = Run(experiment_name=...)
    run.log_configs(
        {
            "parameters/learning_rate": 0.001,
            "parameters/batch_size": 64,
        },
    )

Handle nested structures and type casting:

data = {
    "metrics": {
        "token_count": 76420,
        "agg": {
            "loss": None,
            "acc": None,
        }
    },
    "some_list": [1, "test", None],
}

if __name__ == "__main__":
    run = Run(experiment_name=...)

    run.log_configs(
        data=data,
        flatten=True,
        cast_unsupported=True,
    )