Weights and Biases
A plugin that handles logging to a Weights and Biases project.
WandBPlugin logs data and training information to a Weights and Biases project, making it easy to keep track of preprocessing strategies and models, as well as creating experiment reports.
WandBPlugin
WandBPlugin (project:Optional[str]=None, entity:Optional[str]=None, **kwargs)
Weights and Biases plugin. Kwargs from the constructor are passed to wandb.init().
| Type | Default | Details | |
|---|---|---|---|
| project | typing.Optional[str] | None | Name of the Weights and Biases project. |
| entity | typing.Optional[str] | None | Name of the Weights and Biases entity (username). |
| kwargs |
WandBPlugin.on_setup_data
WandBPlugin.on_setup_data ()
Log features and target(s) as a single dataset artifact.
WandBPlugin.on_setup_end
WandBPlugin.on_setup_end ()
Log WandB config that contains the parameters passed to the Poniard estimator constructor.
WandBPlugin.on_infer_types
WandBPlugin.on_infer_types ()
Log the inferred types dataframe as a table.
WandBPlugin.on_setup_preprocessor
WandBPlugin.on_setup_preprocessor ()
Log the preprocessor’s HTML repr.
WandBPlugin.on_plot
WandBPlugin.on_plot (figure:plotly.graph_objs._figure.Figure, name:str)
Log a Plotly plot.
WandBPlugin.on_fit_end
WandBPlugin.on_fit_end ()
Log the results dataframe as a table.
WandBPlugin.on_analyze_estimator
WandBPlugin.on_analyze_estimator (estimator:sklearn.base.BaseEstimator, name:str)
Log fitted estimator as an artifact.