Compocyte.core.tuner

Classes

Tuner(database_path, adata_path, hierarchy, ...)

class Compocyte.core.tuner.Tuner(database_path, adata_path, hierarchy, root_node, obs_names)

Bases: object

Parameters:

• database_path (str)

• adata_path (str)

• hierarchy (dict)

• root_node (str)

• obs_names (list)

__init__(database_path, adata_path, hierarchy, root_node, obs_names)

Parameters:

• database_path (str)

• adata_path (str)

• hierarchy (dict)

• root_node (str)

• obs_names (list)

train_from_tuner(save_path, adata, parallelize=True, max_cells=None, stratify_by=None, processes=None, intermittent_saving=False)

Parameters:

• save_path (str)

• adata (AnnData)

• max_cells (int)

• stratify_by (str)

• processes (int)

Return type:

HierarchicalClassifier

trial_run(cv_key, n_features, hidden_layers, dropout, epochs, batch_size, starting_lr, max_lr, momentum, beta, gamma, test_factor, parallelize=True, num_threads=None, standardize_separately=None)

Parameters:

• cv_key (str)

• n_features (int)

• hidden_layers (list)

• dropout (float)

• epochs (int)

• batch_size (int)

• starting_lr (float)

• max_lr (float)

• momentum (float)

• beta (float)

• gamma (float)

• test_factor (int)

• parallelize (bool)

• standardize_separately (str)

Return type:

None

make_db()

Return type:

None

make_entry(node, trials, fraction_correct, fraction_matches, n_features, hidden_layers, dropout, epochs, batch_size, starting_lr, max_lr, momentum, beta, gamma, threshold)

Parameters:

• node (str)

• trials (int)

• fraction_correct (float)

• fraction_matches (int)

• n_features (int)

• hidden_layers (str)

• dropout (float)

• epochs (int)

• batch_size (int)

• starting_lr (float)

• max_lr (float)

• momentum (float)

• beta (float)

• gamma (float)

• threshold (float)

Return type:

None

get_best_trial(node)

Return type:

dict