embeddings model

[fabianocastello]

Hello. I am trying to use your examples. Thanks for sharing.

In the PDf notebook, we have this gist:

# Use FAISS to index the embeddings. This will allow us to perform a similarity search on the texts using the embeddings.
# https://python.langchain.com/en/latest/modules/indexes/vectorstores/examples/faiss.html
pdfDocSearch = FAISS.from_texts(textTexts, embeddings)

I am getting this error: InvalidRequestError: "Must provide an 'engine' or 'deployment_id' parameter to create a". I am using OPENAI_EMBEDDING_MODEL_NAME = "text-embedding-ada-002 (Version 2)" but I think that this is wrong. I should specify a deployment that is based on a embedding model such as this above, correct?

Please advice.

Thank you.

[tedd-yuan]

Hello. I am trying to use your examples. Thanks for sharing.

In the PDf notebook, we have this gist:

# Use FAISS to index the embeddings. This will allow us to perform a similarity search on the texts using the embeddings.
# https://python.langchain.com/en/latest/modules/indexes/vectorstores/examples/faiss.html
pdfDocSearch = FAISS.from_texts(textTexts, embeddings)

I am getting this error: InvalidRequestError: "Must provide an 'engine' or 'deployment_id' parameter to create a". I am using OPENAI_EMBEDDING_MODEL_NAME = "text-embedding-ada-002 (Version 2)" but I think that this is wrong. I should specify a deployment that is based on a embedding model such as this above, correct?

Please advice.

Thank you.

Got the same error. Any hint? Thanks!

[cplemm]

Passing in model & engine when creating OpenAIEmbeddings as shown below is working for me:

embeddings = OpenAIEmbeddings(openai_api_key=OPENAI_API_KEY, model=OPENAI_EMBEDDING_MODEL_NAME, engine=OPENAI_EMBEDDING_DEPLOYMENT_NAME, client="azure", chunk_size=1)

OPENAI_EMBEDDING_MODEL_NAME is "text-embedding-ada-002" and OPENAI_EMBEDDING_DEPLOYMENT_NAME is the name of my specific model deployment in Azure OpenAI.