We’ve been trying to use Weaviate on our testing environment.
Our embedding queries don’t seem to work for our application as weaviate tries to connect to make a post request to:
http://{baseUrl}/openai/deployments/{deployment_id}/embeddings?api-version=2024-02-01"
The error we are getting is:
dial tcp: lookup {baseUrl} on {…}: server misbehaving’
The embeddings are made using a different baseUrl than the ones we want to use on Query-time. So i have added “X-OpenAI-BaseURL”: {baseUrl} to the headers and the request url seems right. The Api version is set to 2024-02-01
I was wondering if that is the problem to why we are getting this error message and if so would it be possible to change the api version in the headers as well?
Yes, I added those to the collection definition and those worked for vectorizing our documents. but what about when I want to deploy the application and we are using a different base_url when deployed?
So if i want to vector search with “nearText” using a different base_url. Thas base_url should be added in the headers like “X-OpenAI-BaseURL”: {baseUrl} since I am making REST Api calls