def from_config(cls, global_config):
all_metrics = cls(global_config)
for metric_config in all_metrics.metric_config:
subcls = find_subclass_by_name(MLMetric, metric_config.metric_name)
metric_obj = subcls.from_config(global_config, metric_config)
metric_list = None
if subcls in all_subclasses(StepMetric):
metric_list = all_metrics.step_metrics
if subcls in all_subclasses(EndingMetric):
metric_list = all_metrics.ending_metrics
metric_list.append(metric_obj)
return all_metrics
def from_config(cls, global_config):
subcls = find_subclass_by_name(cls, global_config.graph.graph_name)
return subcls.from_config(global_config)
def from_config(cls, global_config):
io_config = global_config.io
subcls = find_subclass_by_name(cls, io_config.io_name)
return subcls.from_config(global_config)
def from_config(cls, global_config, metric_config):
subcls = find_subclass_by_name(cls, metric_config.metric_name)
return subcls.from_config(global_config, metric_config)
def from_config(cls, global_config):
trainer_config = global_config.trainer
subcls = find_subclass_by_name(cls, trainer_config.trainer_name)
return subcls.from_config(global_config)
def from_config(cls, graph_config):
subcls = find_subclass_by_name(cls, graph_config.model_name)
return subcls.from_config(graph_config)
def from_config(cls, global_config):
predictor_config = global_config.predictor
subcls = find_subclass_by_name(cls, predictor_config.predictor_name)
return subcls.from_config(global_config)