Closed hrntsm closed 1 year ago
Hey @hrntsm can you try running this PR in debug and letting me know what the Advice panel says now?
We tested this with gpt-4 but I got the same issue as you with gpt-3-turbo
Hi @clicketyclackety, I tried #3 and it works well.
But I couldn't get any meaningful results; I'm trying with GPT-3.5-turbo, so maybe it's a bit predictable and weak. I have signed up for the GPT-4 waiting list and look forward to testing its capabilities in full. Thank you for your quick response.
Hi @clicketyclackety, I tried #3 and it works well.
But I couldn't get any meaningful results; I'm trying with GPT-3.5-turbo, so maybe it's a bit predictable and weak. I have signed up for the GPT-4 waiting list and look forward to testing its capabilities in full. Thank you for your quick response.
Awesome @hrntsm , I'll try it myself again and release a new build if it works for me too. Yeah it struggles with rectangles. I think it wants the component but currently it creates a parameter. I'll do some GPT3.5/4 comparisons
Error: 429 Too Many Requests { "error": { "message": "You exceeded your current quota, please check your plan and billing details.", "type": "insufficient_quota", "param": null, "code": null } }
First of all, thank your team for developing this wonderful component. I am very excited.
I'm trying it out, but it always gives me an "Value cannot be null. Parameter name: source" in an advise panel. Is there something wrong with my input?
Token and model were set according to README. The model is "gpt-3.5-turbo".
environment