autointent.OptimizationConfig

class autointent.OptimizationConfig(/, **data)

Bases: pydantic.BaseModel

Configuration for the optimization process.

One can use it to customize optimization beyond choosing different preset. Instantiate it and pass to autointent.Pipeline.from_optimization_config().

Parameters:

data (Any)

data_config: autointent.configs.DataConfig

search_space: list[dict[str, Any]]

See tutorial on search space customization.

logging_config: autointent.configs.LoggingConfig

See tutorial on logging configuration.

embedder_config: autointent.configs.EmbedderConfig = None

classmethod validate_embedder_config(v)

Validate and convert embedder config to proper type.

Parameters:

v (Any)

Return type:

autointent.configs.EmbedderConfig

cross_encoder_config: autointent.configs.CrossEncoderConfig

transformer_config: autointent.configs.HFModelConfig

hpo_config: autointent.configs.HPOConfig

seed: pydantic.PositiveInt = 42

classmethod from_preset(preset)

Parameters:

preset (autointent.custom_types.SearchSpacePreset)

Return type:

OptimizationConfig