medl.tools.mmar_converter package

1.0
class ConvertAIAAConfiger(mmar_root: str, wf_config_file_name=None)

Bases: dlmed.utils.wfconf.Configurator

Config parser to load AIAA JSON config of Clara-TF 3.1 and convert it to Clara 4.0.

configure()
finalize_config(config_ctx: dlmed.utils.wfconf.ConfigContext)
get_converted()
process_config_element(config_ctx: dlmed.utils.wfconf.ConfigContext, node: dlmed.utils.json_scanner.Node)
start_config(config_ctx: dlmed.utils.wfconf.ConfigContext)
convert_mmar(mmar_path: str, out_path: str)
convert_shell_script(source_dir: str, dest_dir: str, filename: str, targetname: str, finetune: bool, val: bool, val_ckpt: bool)
main()
class ConvertEvaluateConfiger(init_config: dict, train_ctx: dict, mmar_root: str, wf_config_file_name=None, env_config_file_name=None, log_config_file_name=None, kv_list=None)

Bases: medl.tools.mmar_converter.convert_tf_configer.ConvertTrainConfiger

As the validation config of Clara 4.0 is simiar to the validation part of train config, just extract the validation config from the converted train config, then adjust values based on validate config.

finalize_config(config_ctx: dlmed.utils.wfconf.ConfigContext)
process_first_pass(node: dlmed.utils.json_scanner.Node)
start_config(config_ctx: dlmed.utils.wfconf.ConfigContext)
class ConvertTrainConfiger(mmar_root: str, wf_config_file_name=None, env_config_file_name=None, log_config_file_name=None, kv_list=None)

Bases: dlmed.utils.clara_conf.ClaraConfiger

Config parser to load Clara-TF 3.1 JSON config and convert to Clara 4.0 JSON config.

configure()
finalize_config(config_ctx: dlmed.utils.wfconf.ConfigContext)
get_converted()
process_config_element(config_ctx: dlmed.utils.wfconf.ConfigContext, node: dlmed.utils.json_scanner.Node)
process_first_pass(node: dlmed.utils.json_scanner.Node)
process_second_pass(node: dlmed.utils.json_scanner.Node)
start_config(config_ctx: dlmed.utils.wfconf.ConfigContext)
© Copyright 2021, NVIDIA. Last updated on Feb 2, 2023.