AgentOptimizationPatch.md

AgentOptimizationPatch

Properties

Name	Type	Description	Notes
max_attempts	int	The maximum number of attempts	[optional]
model_choices	List[str]	The models to try	[optional]
judge_model	str	The model used to judge responses	[optional]
variable_choices	List[Dict[str, object]]	Sets of prompt variable values to try	[optional]
acceptance_statements	List[AgentOptimizationAcceptanceStatement]	Statements a response must satisfy to pass	[optional]
judges	List[AgentOptimizationJudge]	Judges that score each response	[optional]
user_input_options	List[str]	User inputs to test with	[optional]
ground_truth_responses	List[str]	Expected responses to compare against	[optional]
metric_key	str	The key of the metric to optimize	[optional]
token_limit	int	The maximum number of tokens to use	[optional]
variation_key	str	The key of the variation to start from	[optional]
label	str	A label for the optimization	[optional]
latency_optimization	bool	Whether to optimize for lower latency	[optional]
token_optimization	bool	Whether to optimize for lower token use	[optional]
auto_commit	bool	Whether to save the best result as a new variation automatically	[optional]

Example

from launchdarkly_api.models.agent_optimization_patch import AgentOptimizationPatch

# TODO update the JSON string below
json = "{}"
# create an instance of AgentOptimizationPatch from a JSON string
agent_optimization_patch_instance = AgentOptimizationPatch.from_json(json)
# print the JSON string representation of the object
print(AgentOptimizationPatch.to_json())

# convert the object into a dict
agent_optimization_patch_dict = agent_optimization_patch_instance.to_dict()
# create an instance of AgentOptimizationPatch from a dict
agent_optimization_patch_from_dict = AgentOptimizationPatch.from_dict(agent_optimization_patch_dict)

[Back to Model list] [Back to API list] [Back to README]