sfaira.train.GridsearchContainer

class sfaira.train.GridsearchContainer(source_path: dict, cv: bool)

Attributes

cv_keys

Returns keys of cross-validation used in dictionaries in this class.

histories

evals

run_ids

gs_keys

summary_tab

cv

source_path

model_id_len

Methods

best_model_by_partition(partition_select, ...)

param partition_select

get_best_model_ids(tab, metric_select, ...)

param tab

load_gs(gs_ids)

Loads all relevant data of a grid search.

load_y(hat_or_true, run_id)

plot_best_model_by_hyperparam(metric_select)

Produces boxplots for all hyperparameters choices by organ.

plot_completions([groupby, height_fig, ...])

Plot number of completed grid search points by category.

plot_training_history(metric_select, metric_show)

Plot train and validation loss during training and learning rate reduction for each organ

save_best_weight(path[, partition, metric, ...])

Copies weight file from best hyperparameter setting from grid search directory to zoo directory with cleaned file name.

write_best_hyparam(write_path[, subset, ...])