Skip to content

ExperimentConfig

Defines the database and schema that experiment metrics, experiment metadata and registered models will be saved.

⚠️ Note:

  • During each training run (a single call to .fit()), the model with the best validation metric across all epochs is saved.
  • Ensure the native app has the necessary permissions to the specified database and schema.
-- grant access to resources needed for snowflake experiment tracking
GRANT USAGE ON DATABASE <DATABASE> TO APPLICATION RELATIONALAI;
GRANT USAGE ON SCHEMA <DATABASE>.<SCHEMA> TO APPLICATION RELATIONALAI;
GRANT CREATE EXPERIMENT ON SCHEMA <DATABASE>.<SCHEMA> TO APPLICATION RELATIONALAI;
ParameterTypeDescription
DatabasestrDatabase to save the experiment metadata.
SchemastrSchema in database to save experiment metadata.

An instance of the ExperimentConfig class.

from relationalai_gnns import ExperimentConfig
# Create an instance of ExperimentConfig; artifacts will be saved in "database_name.schema_name"
experiment_cfg = ExperimentConfig(database="database_name", schema="schema_name")

Returns a dictionary representation of the ExperimentConfig instance.

Example:

experiment_cfg.to_dict()