-
Notifications
You must be signed in to change notification settings - Fork 16
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 gettig langchain API connectionError retries #51
Comments
Hi @DavidCurryHan, Based on the info provided, this seems like a network connectivity or proxy issue. During training, Rasa uses only the embeddings API, not the completions API, which might be blocked. |
Thank you for your reply! |
as shown in the image
I have followed the instruction and successfully train the rasa model and actived the GUI interface.
However as long as i type anything, the CALM-demo keep giving me this openai APIconnectionError shown in yellow.
Since I have successfully train the model, the program should have accessed the openai with no problem, it should not be a API key problem.
Also, I have write my own script of "OPENAI.chat.completions.create" which also works fine with the openai library version above 1.0.0.(since this rasa framework is using openai version 0.28.1)
I have installed the openai back to 0.28.1 and tried to set the timeout to 100seconds. I have also tried to set my proxy port to localhost:7890 and OPENAI_API_BASE to "https://openai.wndbac.cn/v1".
Eventually, the error stays.
Im not sure if its an openai problem (since the model training worked at 0.28.1 and script worked at 1.0.0) or its a problem from the rasa end.
Can anyone help? thankyou!
The text was updated successfully, but these errors were encountered: