Will chatglm3-6b support parallel-function-calling ?

#53
by simonwei97 - opened

Will chatglm3-6b support parallel-function-calling ?

Refer: https://platform.openai.com/docs/guides/function-calling

Now

image.png

What I Expected

Q: What's the weather like in San Francisco, Tokyo, and Paris?

chatglm3-6b output should be as following:

# LLM response
tool_call: [
  {'name': 'get_current_weather', 'parameters': {'location': 'San Francisco'}},
  {'name': 'get_current_weather', 'parameters': {'location': 'Tokyo'}},
  {'name': 'get_current_weather', 'parameters': {'location': 'Paris'}}
]

Then User send the info for each function call and function response to the model. Do not need call 3 times LLM.

More example like this.

https://docs.llamaindex.ai/en/stable/examples/agent/openai_agent_parallel_function_calling/#example-from-openai-docs

different gpt3.5 version

llm = OpenAI(model="gpt-3.5-turbo-0613")
agent = OpenAIAgent.from_tools([weather_tool], llm=llm, verbose=True)
response = agent.chat(
    "What's the weather like in San Francisco, Tokyo, and Paris?"
)
  • Issue will be will be solved within a single turn of dialogue for gpt-3.5-turbo-1106. (Will chatglm3-6b run like this?)
  • Issue will be will be solved within 3 separate turns for gpt-3.5-turbo-0613, which is not quite advanced.
Knowledge Engineering Group (KEG) & Data Mining at Tsinghua University org

no support

zRzRzRzRzRzRzR changed discussion status to closed

Sign up or log in to comment