bruvduroiu
commented
4 months ago Hi @hardcorebadger , thank you for raising this. We do something similar when using local models, by constraining output using GBNF.
For functions fine-tuning, is this example reflective of what you're referring to, where function signature is being passed to OpenAI as a Tool?
# Example dummy function hard coded to return the same weather
# In production, this could be your backend API or an external API
def get_current_weather(location, unit="fahrenheit"):
"""Get the current weather in a given location"""
if "tokyo" in location.lower():
return json.dumps({"location": "Tokyo", "temperature": "10", "unit": unit})
elif "san francisco" in location.lower():
return json.dumps({"location": "San Francisco", "temperature": "72", "unit": unit})
elif "paris" in location.lower():
return json.dumps({"location": "Paris", "temperature": "22", "unit": unit})
else:
return json.dumps({"location": location, "temperature": "unknown"})
def run_conversation():
# Step 1: send the conversation and available functions to the model
messages = [{"role": "user", "content": "What's the weather like in San Francisco, Tokyo, and Paris?"}]
tools = [
{
"type": "function",
"function": {
"name": "get_current_weather",
"description": "Get the current weather in a given location",
"parameters": {
"type": "object",
"properties": {
"location": {
"type": "string",
"description": "The city and state, e.g. San Francisco, CA",
},
"unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
},
"required": ["location"],
},
},
}
]
response = client.chat.completions.create(
model="gpt-3.5-turbo-0125",
messages=messages,
tools=tools,
tool_choice="auto", # auto is default, but we'll be explicit
)
Siraj-Aizlewood
hardcorebadger
Love the concept, but ran into issues with param parsing on dynamic routes with openAI, even with simple functions, when compared to running a completion on their functions fine tuning. There's likely an optimal way to parse function params using openAI's fine tuning and format, rather than having to prompt engineer 3.5-turbo into a JSON function call.
Obviously it's a bit less agnostic to foundational model, but could be dealt with via inheritance, and would make the "easy mode" work a lot better out of the box :)