Continue generating on stop_reason: length

[34r7h]

Describe the feature or improvement you're requesting

Is there a way to continue generating without backfeeding the entire message everytime?

Additional context

Atm my code is doing this which I feel is very wasteful

import sys
import json
import openai
from dotenv import load_dotenv
import os

def recursive_chat(messages):
    chat_completion = openai.ChatCompletion.create(
        model="gpt-3.5-turbo",
        messages=messages,
        max_tokens=140
    )
    response_message = chat_completion['choices'][0]['message']
    print(response_message)
    messages[0]['content'] += ' ' + response_message['content']  # append the assistant's message to the user's message

    if chat_completion['choices'][0]['finish_reason'] == 'stop':
        return messages
    else:
        return recursive_chat(messages)

# load .env file
load_dotenv()
print(sys.argv)
prompt = sys.argv[1] if len(sys.argv) > 1 else "2+2"
key = sys.argv[2] if len(sys.argv) > 2 else os.getenv('OPENAI_API_KEY')
# Load your API key from an environment variable or secret management service
openai.api_key = key

messages = [{"role": "user", "content": "how could openai fail?"}]

final_messages = recursive_chat(messages)
print(json.dumps(final_messages))

[logankilpatrick]

No, right now you need to pass all the previous context if you want to continue generating, you can write some simple code to do this, my truncating the previous messages, having the model summarize them, and then asking it to continue, but out of the box we don't support this.