nat.plugins.openpipe.config

Classes

ARTTrajectoryBuilderConfig	Configuration for the OpenPipe ART Trajectory Builder.
ARTBackendConfig	Base configuration for the ART backend.
ARTTrainerAdapterConfig	Configuration for the ART Trainer run
ARTTrainerConfig	Configuration for the ART Trainer run

Module Contents

class ARTTrajectoryBuilderConfig(/, **data: Any)

Bases: nat.data_models.finetuning.TrajectoryBuilderConfig

Configuration for the OpenPipe ART Trajectory Builder.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

num_generations: int = None

class ARTBackendConfig(/, **data: Any)

Bases: pydantic.BaseModel

Base configuration for the ART backend.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

ip: str = None

port: int = None

name: str = None

project: str = None

base_model: str = None

api_key: str = None

delete_old_checkpoints: bool = None

init_args: art.dev.InitArgs | None = None

engine_args: art.dev.EngineArgs | None = None

torchtune_args: art.dev.TorchtuneArgs | None = None

server_config: art.dev.OpenAIServerConfig | None = None

class ARTTrainerAdapterConfig(/, **data: Any)

Bases: nat.data_models.finetuning.TrainerAdapterConfig

Configuration for the ART Trainer run

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

backend: ARTBackendConfig = None

training: art.dev.TrainerArgs | None = None

class ARTTrainerConfig(/, **data: Any)

Bases: nat.data_models.finetuning.TrainerConfig

Configuration for the ART Trainer run

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.