Skip to content

Latest commit

 

History

History
122 lines (97 loc) · 5.8 KB

6_2_1_Discussion_Group.md

File metadata and controls

122 lines (97 loc) · 5.8 KB
Raw

Discussion: Multi-Agent Discussion Group

In this use case, we will demonstrate how to use agentUniverse for multi-agent discussions.

There are two agent roles: discussion group host (one) and discussion group participants (several).

The user initiates a discussion topic, and the host organizes the participants to start the discussion. In each round, each participant expresses his or her own views according to the topic of the discussion (the views of each participant are not immutable and will be constantly adjusted as the discussion goes on). After several rounds of discussion, the host will summarize the discussion process and return the results of the participants after several rounds of discussion to the user.

Many heads are better than one, imagine that you have a question you want to answer and submit it directly to the agentUniverse discussion group, The host will organize multiple agents with different ideas together, focus on your question, and finally give you a comprehensive and intelligent answer, which is an interesting agent experiment.

Quick Start

Configure API Key

For example, configure key information in the file custom_ key.toml in which agentUniverse manages private key configuration (The agentUniverse discussion group uses GPT as the base model and Serper as the Google search tool by default. The following describes how to use other models or tools).

[KEY_LIST]
# serper google search key
SERPER_API_KEY='xxx'
# openai api key
OPENAI_API_KEY='xxx'

Modify Prompt Version

Modify the prompt version of agents to the english version, the discussion group multi-agent configuration file is located in the app/core/agent/discussion_agent_case directory in the sample_standard_app sample project. Find the host_agent.yaml and change prompt_version to host_agent.en, similarly find the participants yaml file and change the prompt_version to participant_agent.en.

info:
  name: 'host_agent'
  description: 'host agent'
profile:
  prompt_version: host_agent.en
  llm_model:
    name: 'demo_llm'
    temperature: 0.6
plan:
  planner:
    name: 'discussion_planner'
    round: 2
    participant:
      name:
        - 'participant_agent_one'
        - 'participant_agent_two'
memory:
  name: 'demo_memory'
metadata:
  type: 'AGENT'
  module: 'sample_standard_app.app.core.agent.discussion_agent_case.host_agent'
  class: 'HostAgent'

Run Discussion Group

In the sample project of agentUnvierse sample_standard_app, find the discussion_chat_bots.py file in the app/examples directory, enter the question you want to answer in the chat method, and run it.

For example, enter the question: Which tastes better, Coca-Cola or Pepsi

from agentuniverse.base.agentuniverse import AgentUniverse
from agentuniverse.agent.agent import Agent
from agentuniverse.agent.agent_manager import AgentManager

AgentUniverse().start(config_path='../../config/config.toml')


def chat(question: str):
    instance: Agent = AgentManager().get_instance_obj('host_agent')
    instance.run(input=question)


if __name__ == '__main__':
    chat("Which tastes better, Coca-Cola or Pepsi")

Result Demonstration

Which tastes better, Coca-Cola or Pepsi：

Picture

More Details

Agent Configuration

The discussion group multi-agent configuration files are located in the app/core/agent/discussion_agent_case directory in the sample_standard_app sample project.

The host corresponds to host_agent.yaml and the participant corresponds to participant_agent_*.yaml.

The prompt is managed by the prompt version. For example, the default prompt_version of host_agent is host_agent.cn, and the corresponding prompt file is host_agent.cn.yaml under the app/core/prompt directory.

If the LLM Model will be modified to another, such as the Qwen model, change the llm_model name information (such as the qwen_llm built into the aU system).

The number of discussion rounds in agent plan defaults to 2, which users can adjust as needed. The default participants are two agents built into aU, which can add discussion group members by creating new agents and configuring the agent name parameter to the participant of the planner.

info:
  name: 'host_agent'
  description: 'host agent'
profile:
  prompt_version: host_agent.cn
  llm_model:
    name: 'qwen_llm'
    temperature: 0.6
plan:
  planner:
    name: 'discussion_planner'
    round: 2
    participant:
      name:
        - 'participant_agent_one'
        - 'participant_agent_two'
memory:
  name: 'demo_memory'
metadata:
  type: 'AGENT'
  module: 'sample_standard_app.app.core.agent.discussion_agent_case.host_agent'
  class: 'HostAgent'

Memory Configuration

Participants in the aU discussion group configure demo_memory in the sample_standard_app sample project by default to store common memory information for the entire discussion group.

Planner Configuration

The planner configuration file is located in the app/core/planner directory of the sample_standard_app sample project, where the discussion_planner.py file is the specific plan process code of the discussion group. If you are interested, you can read it yourself.

Prompt Configuration

The prompt is managed by the prompt version, and the default prompt files of the aU discussion group agent are configured in the app/core/prompt directory. Corresponding to the Chinese and English versions (host_agent_xx.yaml/participant_agent_xx.yaml), users can change the prompt_version in the agent configuration to complete fast switching.

Tool Configuration

Two participants in the aU discussion group are configured with google_search_tool by default. Users can change the tool name in the participant_agent_xxx.yaml under the app/core/agent/discussion_agent_case directory in the sample_standard_app sample project to switch tool calls.