briansunter/logseq-plugin-gpt3-openai

Doesn't work with chatGPT

Arbitrate3280 opened this issue ยท 10 comments

OpenAI just released the chatgpt api:
https://openai.com/blog/introducing-chatgpt-and-whisper-apis

https://platform.openai.com/docs/models/gpt-3-5

I've added "gpt-3.5-turbo" in openAICompletionEngine in the plugin settings but it doesn't work.

I get this:
Unknown OpenAI Error

and

image

Hey @Vitorvlv , I'm working on supporting this now. Should be out in a couple days.

This new model requires me to call a new endpoint, unlike the previous models they released, so it takes some changes on my side, but should be pretty straightforward to implement.

Should be a nice update since they say it's as good as the earlier models but 10x cheaper!

Hey @Vitorvlv , I'm working on supporting this now. Should be out in a couple days.

This new model requires me to call a new endpoint, unlike the previous models they released, so it takes some changes on my side, but should be pretty straightforward to implement.

Should be a nice update since they say it's as good as the earlier models but 10x cheaper!

Great! This is the first thing I tried when I got the email from Open AI this morning: your plug-in is super useful! Thanks a lot for that!
In the same vein, as the Whisper API is now available, being able to upload an audio file and get the transcribed text directly in Logeq would be pretty cool!

@marcsachet I'll definitely implement the whisper API integration as well. It would be really cool to take voice notes using the app and then use the plugin to transcribe them

I updated the project roadmap to add that. https://github.com/users/briansunter/projects/1

Right now, I plan on implementing this, then the multi-block select, then the whisper integration.

Should be able to finish the chatgpt pretty quick, then get whisper out in the next couple of weeks.

@marcsachet I'll definitely implement the whisper API integration as well. It would be really cool to take voice notes using the app and then use the plugin to transcribe them

I updated the project roadmap to add that. https://github.com/users/briansunter/projects/1

Right now, I plan on implementing this, then the multi-block select, then the whisper integration.

Should be able to finish the chatgpt pretty quick, then get whisper out in the next couple of weeks.

Fantastic! it'll be very cool indeed! and not to mention the possible (almost) on the fly translation!

just gave it a shot, working great, thanks!!

Awesome! Glad you like it!

Thanks, @briansunter This very quick work, is Awesome.

Still not working for me
**
image
**

What's the solution for this issue "Unknown OpenAI Error"?

Still not working for me ** image **

First, try to change/regenerate your API key for logseq.
Second, make sure the number "OpenAI Max Tokens" in the plugin setting is 2048 or lower. It fixed my problem with this.

Still not working for me ** image **

same, doesn't work