sfaira.train.GridsearchContainer.plot_best_model_by_hyperparam¶
- GridsearchContainer.plot_best_model_by_hyperparam(metric_select: str, param_hue='lr', partition_select: str = 'val', partition_show: str = 'test', subset: dict = {}, param_x: Union[tuple, list] = ('lr', 'depth', 'width', 'dropout', 'l1', 'l2'), show_swarm: bool = False, panel_width: float = 4.0, panel_height: float = 2.0)¶
Produces boxplots for all hyperparameters choices by organ.
- Parameters
partition – “train” or “eval” or “test” partition of data.
metric_select – Metric to plot.
param_x – Hyper-parameter for x-axis partition.
param_hue – Hyper-parameter for hue-axis partition.
panel_width –
panel_height –
- Returns