Llama 3.1 原生 Function Calling 深度体验（LMDeploy 版）
输入“/”快速插入内容
Llama 3.1 原生 Function Calling 深度体验（LMDeploy 版）
Llama 3.1 原生 Function Calling 深度体验（LMDeploy 版）
用户1672
用户95832024年7月24日修改
Meta 发布了 Llama 3.1 第一次支持了原生 Function Calling 功能，本文将带大家使用 LMDeploy 深度体验 Llama 3.1 的 Function Calling 功能。
InternLM2.5 Function Calling
代码块
from openai import OpenAI
import math
import addict
from func_timeout import func_timeout
def _safe_eval(expr):
math_methods = {k: v for k, v in math.__dict__.items()
if not k.startswith('_')}
allowed_methods = {
'math': addict.Addict(math_methods),
'max': max,
'min': min,
'round': round,
'sum': sum,
**math_methods,
}
allowed_methods['__builtins__'] = None
return eval(expr, allowed_methods, allowed_methods)
def calculator(expression: str):
res = func_timeout(timeout=5, func=_safe_eval, args=[expression])
return f'The result of {expression} is {res}'
tools = [
{
"type": "function",
"function": {
"name": "calculator",
"description": ('A calculator tool. The input must be a single Python '
'expression and you cannot import packages. You can use functions '
'in the `math` package without import.'),
"parameters": {
"type": "object",
"properties": {
"expression": {
"type": "string",
"description": "The expression for calculator",
},
},
"required": ["expression"],
},
}
}
]
client = OpenAI(api_key='YOUR_API_KEY', base_url='http://0.0.0.0:23333/v1')
model_name = client.models.list().data[0].id
messages = [{'role': 'user', 'content': '使用计算器，13.8 和 13.11 谁更大？'}]
response = client.chat.completions.create(
model=model_name,
messages=messages,
temperature=0.8,
top_p=0.8,
stream=False,
tools=tools)
print(response)
func1_name = response.choices[0].message.tool_calls[0].function.name
func1_args = response.choices[0].message.tool_calls[0].function.arguments
func1_out = eval(f'{func1_name}(**{func1_args})')
print(func1_out)
messages.append(
{'role': 'assistant', 'content': response.choices[0].message.content})
messages.append(
{'role': 'environment', 'content': func1_out, 'name': 'plugin'})
response = client.chat.completions.create(
model=model_name,
messages=messages,
temperature=0.8,
top_p=0.8,
stream=False,
tools=tools)
print(response)
messages.append(
{'role': 'assistant', 'content': response.choices[0].message.content})
for message in messages:
print(f'{message["role"]}: {message["content"]}\n')