LightGBM
When using auto_record()
with LightGBM models, the following information is saved:
- Hyper-parameters: Details about how the model is configured, like settings and options.
- Metrics: Records epoch-time versus epoch and any custom metrics logged by the user. If no metrics are specified, LightGBM logs loss on the specified validation sets by default.
To use auto_record
with LighGBM in MarkovML, provide the following:
name
: LighGBM model name.notes
: Notes for future reference. (optional)project_id
: Project ID of the Project you will work on.model_class
: Define Markov model class.
Sample Code
import markov
MODEL_NAME = "My Test LighGBM Model"
markov.lightgbm.auto_record(
name=MODEL_NAME,
notes="Testing LighGBM Auto Record with MarkovML",
project_id="some_project_id",
model_class=markov.ModelClass.CLASSIFICATION
)
# Continue creation LighGBM model and training
Updated about 1 month ago