-
Notifications
You must be signed in to change notification settings - Fork 1k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Keep hitting the timeout #29
Comments
thanks for the help! this is very strange. calling in friends from modal |
hey @sanxfxteam - we had a look and think it was a transient network issue on your end or on Modal's end or on OpenAI's end. we reallyhave no way of knowing. @gerred reran everything and couldnt repro. please let reopen this issue if the problem persists and send us your prompt! |
I suspect it might be this. Would be nice to have a client-side timeout for the chat completion. I can try and add that in a little bit :) |
I was using GPT-4 model, maybe the problem was there too. Will try again
later. Thanks for the quick response.
…On Sun, May 21, 2023, 17:27 Akshat Bubna ***@***.***> wrote:
I suspect it might be this
<https://community.openai.com/t/completion-endpoint-randomly-freezes/107206/4>.
Would be nice to have a client-side timeout for the chat completion. I can
try and add that in a little bit :)
—
Reply to this email directly, view it on GitHub
<#29 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AADNGNZVGCXSUXLOZYX4RVDXHKXHLANCNFSM6AAAAAAYJALIOI>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
@sanxfxteam If you run into it again and it's repeatable for a given prompt, send it over and I'm happy to investigate and provide some guidance. At least over in local LLM land with very low temperatures I can repro issues where models either get stuck in repeat loops or don't provide parseable tokens to move forward, and these issues can still happen with SOTA models. Don't think this is a smol or Modal issue. |
src/templates/ai_recommendation.html {% block content %} AI-Powered Prediction & RecommendationDiscover new Pokemon based on your preferences and search history.
Get Recommendations
|
same error |
this is my prompt Title: PokeFinder App Description: Discover, track and collect detailed information about your favorite Pokemon with the PokeFinder App. Leverage the power of artificial intelligence to find and learn about the Pokemon you love. Objective: The objective of the PokeFinder App is to help Pokemon enthusiasts explore, discover, and learn about their favorite Pokemon characters. Users can search and track these creatures, collect detailed information, and enjoy an engaging, fun experience as a Pokemon fan. Key Features:
Additional Requirements:
|
I was having the image issue where it would crash beacause it cant tell if it needs to create images, |
now I am getting 300 sec rate limit error |
I also repeatedly run into this when trying to generate an app. I'm using gpt-4. |
I had this with a large fairly complicated prompt and I figured this out. you have to update the main.py file and change @stub.function() to @stub.function(timeout=30000) This is for the API call to modal to update the timeout for the full program build length and that seems to have done it for my api calls hitting limits. Hope it helps you guys! |
thanks @JeffreyBoehme ! |
I've change things in main.py:
I use modal.
The text was updated successfully, but these errors were encountered: