nip.scenario_base.environment

nip.scenario_base.environment#

Base class for the RL environment.

Classes

Environment(hyper_params, settings, dataset, ...)

The base class for all Prover-Verifier RL environments.

PromptMessage

A message in the prompt for a language model API.

PureTextEnvironment(hyper_params, settings, ...)

Base for environments which handle non-tokenised text with nested array dicts.

TensorDictEnvironment(*args, **kwargs)

The base class for all Prover-Verifier RL environments which use tensordicts.