Skip to main content

ModelDraft.train_async

ModelDraft.train(engine=None)

Asynchronously train a model from the model draft with the current config and dataset
that have been set.

Parameters:

      engine: Engine, default None
The engine to use for this training job.

Returns:

      Model

Examples:

Get an already created model by name and get a new draft for iteration, then set new config on draft.

import yaml

with open("~/{path to file}/improved_model_config.yaml", "r") as f:
config = yaml.safe_load(f)

model = pc.get_model("Titanic Model")
draft = model.to_draft()
draft.config = config
model = draft.train_async()