cluster.automl package

Submodules

cluster.automl.automl_runmanager module

class cluster.automl.automl_runmanager.AutoMlRunManager(nn_id)

Bases: object

automl runmanager with handles from create net ver, batch ver

create_networks(generation, number)

We need to create new network for evluate our hyperparameter :param generation:number of generation :param population:number of population for each generation :return:return list of network (nn_id and version)

create_version(generation)

create version on database :param generation: :return:

discard_inferior(networks, survive)

discard inferior genes combinations :param networks: network lists :return: networks

get_all_nodes_list(nn_id, wf_ver)

get execute class path :param node_id: :return:

mutation(networks, number)

mutate gene with survived ones :param networks: networks info :param number: number of genes to generate :return: networks

run()

run automl :return:

save_summary(info)

save best survived results :return:

set_train_finish(ver_data_sets)

update state flag for version level info on database :param input_data : raw info we had on database :return:version info

set_value(data_set, key, value)

train_networks(networks)

train each networks on cluster server :param networks: network lists :return: networks

update_summary(networks, survive)

update summary info of processing genetic algorithm arranged by generation (sequece of list means number of generation each node contains extra info like survive or not :param networks: networks for one generations :param survive: number of gene to survive :return: dict type result info with extra flag

cluster.automl.automl_runmanager.automl_run(nnid)

Module contents