prompting.validators.reward.config

Module Contents

Classes

RewardModelType	Create a collection of name/value pairs.
DefaultRewardFrameworkConfig	Reward framework default configuration.

class prompting.validators.reward.config.RewardModelType(*args, **kwds)

Bases: enum.Enum

Create a collection of name/value pairs.

Example enumeration:

>>> class Color(Enum):
...     RED = 1
...     BLUE = 2
...     GREEN = 3

Access them by:

• attribute access:

  >>> Color.RED
  <Color.RED: 1>
  

• value lookup:

  >>> Color(1)
  <Color.RED: 1>
  

• name lookup:

  >>> Color['RED']
  <Color.RED: 1>
  

Enumerations can be iterated over, and know how many members they have:

>>> len(Color)
3

>>> list(Color)
[<Color.RED: 1>, <Color.BLUE: 2>, <Color.GREEN: 3>]

Methods can be added to enumerations, and members can have their own attributes – see the documentation for details.

dpo = 'dpo_reward_model'

rlhf = 'rlhf_reward_model'

reciprocate = 'reciprocate_reward_model'

dahoas = 'dahoas_reward_model'

diversity = 'diversity_reward_model'

prompt = 'prompt_reward_model'

blacklist = 'blacklist_filter'

nsfw = 'nsfw_filter'

relevance = 'relevance_filter'

relevance_bert = 'relevance_bert'

relevance_mpnet = 'relevance_mpnet'

task_validator = 'task_validator_filter'

keyword_match = 'keyword_match_penalty'

class prompting.validators.reward.config.DefaultRewardFrameworkConfig

Reward framework default configuration. Note: All the weights should add up to 1.0.

dpo_model_weight: float = 0.6

rlhf_model_weight: float = 0

reciprocate_model_weight: float = 0.4

dahoas_model_weight: float = 0

prompt_model_weight: float = 0