No deployment found for OpenAI Embedding

[usinha3]

I have a foundation model gpt-4o deployed.

However when using text-embedding-ada-002 for RAG vector embedding, I keep running into the issue 

ValueError: No deployment found with: deployment.model_name == text-embedding-ada-002

I have an orchestration deployed as well, but still the same issue. 

Any help will be appreciated

[usinha3]

I have a foundation model gpt-4o deployed.

However when using text-embedding-ada-002 for RAG vector embedding, I keep running into the issue 

ValueError: No deployment found with: deployment.model_name == text-embedding-ada-002

I have an orchestration deployed as well, but still the same issue. 

Any help will be appreciated

[DKK]

Hi,

Unfortunately, it is no supported today. Please choose either "Text Embedding 3 Large" or "Text Embedding 3 Small",

Thanks

[usinha3]

Hi, I still get the same error even with the new embedding model. 

For a simple function like this, 

I keep getting 

Regardless of the model used. 

[DKK]

Hi, can you try the following?

from gen_ai_hub.proxy.langchain.init_models import init_embedding_model
from Parameters import * #define the connection to your aicore
import os

os.environ["AICORE_CLIENT_ID"] = AICORE_CLIENT_ID
os.environ["AICORE_CLIENT_SECRET"] = AICORE_CLIENT_SECRET
os.environ["AICORE_AUTH_URL"] = AICORE_AUTH_URL
os.environ["AICORE_RESOURCE_GROUP"] = AICORE_RESOURCE_GROUP
os.environ["AICORE_BASE_URL"] = AICORE_BASE_URL

embeddings = init_embedding_model('text-embedding-3-large')
embeddings

If this is now working, try please 

pip install "generative-ai-hub-sdk[all]"

[mcodricis]

HI, the Deployment ID error appears for any model when running LLM requests through the SAP SDK or orchestration. Supplying the Deployment ID at a low-level request prevents the error. In the generative-ai-hub-sdk package, the current deployment is no longer determined automatically, and the methods do not include a parameter to specify it.

[Vitaliy-R]

Have you deployed any of the embedding models? Because in your post, you mentioned only deploying gpt-4o.