search

modulersYJ / ganoverflow-front

2 stars 2 forks source link

feat-chat: parital answer 해결을 위해 토큰 수 제한 해제 #161

Closed ABizCho closed 9 months ago

ABizCho commented 9 months ago

Partial Answer 문제를 해결하기 위해 아래와 같이 max_tokens를 3000에서 3.5 turbo의 최대치인 4096으로 늘렸습니다.


  const payload: OpenAIStreamPayload = {
    model: "gpt-3.5-turbo",
    messages: prompts,
    temperature: 0.7,
    top_p: 1,
    frequency_penalty: 0,
    presence_penalty: 0,
    max_tokens: 4096,
    stream: true,
    n: 1,
  };
vercel[bot] commented 9 months ago

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name Status Preview Comments Updated (UTC)
ganoverflow ✅ Ready (Inspect) Visit Preview ✅ 1 resolved Jan 10, 2024 1:05pm
ganoverflow-front ✅ Ready (Inspect) Visit Preview 💬 Add feedback Jan 10, 2024 1:05pm