agents.reasoning_and_critique.tot.orchestratorΒΆ
Tree of Thoughts Orchestrator using MultiAgent.
This module implements the Tree of Thoughts algorithm using a multi-agent approach with MultiAgent coordinating the Candidate Generator and Solution Scorer agents.
ClassesΒΆ
Result from Tree of Thoughts execution. |
|
Orchestrates Tree of Thoughts algorithm using multiple agents. |
FunctionsΒΆ
|
Factory function to create a Tree of Thoughts solver. |
Module ContentsΒΆ
- class agents.reasoning_and_critique.tot.orchestrator.TOTResult(/, **data)ΒΆ
Bases:
pydantic.BaseModel
Result from Tree of Thoughts execution.
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.
- Parameters:
data (Any)
- class agents.reasoning_and_critique.tot.orchestrator.TreeOfThoughtsOrchestrator(name='tot_orchestrator', engine=None, beam_width=5, max_iterations=3, temperature_generate=0.7, temperature_score=0.3)ΒΆ
Orchestrates Tree of Thoughts algorithm using multiple agents.
Initialize the Tree of Thoughts orchestrator.
- Parameters:
name (str) β Name for the orchestrator
engine (haive.core.engine.aug_llm.AugLLMConfig | None) β LLM configuration for the coordinator
beam_width (int) β Number of top solutions to keep at each iteration
max_iterations (int) β Maximum number of expansion iterations
temperature_generate (float) β Temperature for candidate generation
temperature_score (float) β Temperature for solution scoring
- async solve(problem, initial_seed=None, context='')ΒΆ
Solve a problem using Tree of Thoughts.
- async agents.reasoning_and_critique.tot.orchestrator.create_tot_solver(beam_width=5, max_iterations=3, **kwargs)ΒΆ
Factory function to create a Tree of Thoughts solver.
- Parameters:
- Returns:
Configured TreeOfThoughtsOrchestrator instance
- Return type: