microsoft / PubSec-Info-Assistant

Information Assistant, built with Azure OpenAI Service, Industry Accelerator

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Gov deployment embedding error

nhwkuhns opened this issue · comments

Bug Details

Describe the bug
When deploying to Azure Gov with the IS_USGOV_DEPLOYMENT=true flag with the ada embedding model, the environment variable in the enrichment web-app gets set to True which causes a 500 error to appear on the front-end when prompting the system as well as error out any document enrichments. I will add a method to ensure the text is all lower-case in the enrichment web-app python code.

Steps To Reproduce

  1. Deploy to Azure Gov
  2. Select Azure OpenAI Ada as the embedding service
  3. ask any prompt to the frontend

GitHub branch: [e.g. main]
main
Version or Latest commit: [obtained by running git log -n 1 <branchname>
43cb0f0 637beb5
What region is your Azure Open AI Service in?
Azure Gov
What ChatGPT model are you using?
gpt-4
model name: (i.e. gpt-3.5-turbo, gpt-4)
gpt-4
model version: (i.e. 0613)
0125-Preview
What embeddings model are you using?
text-embedding-ada-002

If the bug is confirmed, would you be willing to submit a PR?

  • Yes
  • No