[dask] Support init_model

## Summary

LightGBM allows you to provide initial model for continued training.

This option should be supported in the Dask interface.

## Motivation

This change would bring the Dask interface closer to full feature parity with the non-Dask interface, so that users who'd otherwise like to use Dask don't have to avoid it because `init_model` is missing.

## References

https://github.com/microsoft/LightGBM/blob/e5c3f7e755659c120b94c109b515569fa6765b9c/python-package/lightgbm/sklearn.py#L252-L253

https://github.com/microsoft/LightGBM/blob/d6ebd063fff7ff9ed557c3f2bcacc8f9456583e6/python-package/lightgbm/engine.py#L159-L164

https://github.com/microsoft/LightGBM/blob/296397df7bf584770a0297d84ce6a8f4ef25c317/python-package/lightgbm/dask.py#L858

	if isinstance(init_model, str):
	predictor = _InnerPredictor(model_file=init_model, pred_parameter=params)
	elif isinstance(init_model, Booster):
	predictor = init_model._to_predictor(dict(init_model.params, **params))
	else:
	predictor = None

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[dask] Support init_model #4063

StrikerRUS
openedon Mar 12, 2021

Summary

Motivation

References

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

	init_model : string, Booster, LGBMModel or None, optional (default=None)
	Filename of LightGBM model, Booster instance or LGBMModel instance used for continue training.