CPD5.0.3 - watsonx.ai and langchian program error - model_no_support_for function.

[doyoungim999]

Hello,
I am testing small python with langchain and CPD5.0.3.
It looks like embeddings is not supported langchain_ibm.
Is there any other way to run simple application with watsonx.ai ?
I am using cpd5.0.3 in my private environment.

ApiRequestFailure: Failure during generate. (POST https://cp53...../ml/v1/text/embeddings?version=2024-11-27)
Status code: 400, body: {"errors":[{"code":"model_no_support_for_function","message":"Model 'meta-llama/llama-3-1-8b-instruct' does not support function 'function_embedding'","more_info":"https://cloud.ibm.com/apidocs/watsonx-ai-cp"}],"trace":"0f9f152b-19d7-4165-bcc1-7ad75fa35e4e","status_code":400}

---

from langchain_core.vectorstores import InMemoryVectorStore

text = "LangChain is the framework for building context-aware reasoning applications"

vectorstore = InMemoryVectorStore(embeddings)

from langchain_core.documents import Document
document_1 = Document(id="1", page_content="foo", metadata={"baz": "bar"})
document_2 = Document(id="2", page_content="thud", metadata={"bar": "baz"})
document_3 = Document(id="3", page_content="i will be deleted :(")

documents = [document_1, document_2, document_3]
vectorstore.add_documents(documents=documents)

#vectorstore = InMemoryVectorStore.from_texts(

[text],

embedding=embeddings,

#)

Use the vectorstore as a retriever

retriever = vectorstore.as_retriever()

Retrieve the most similar text

retrieved_documents = retriever.invoke("What is LangChain?")

show the retrieved document's content

retrieved_documents[0].page_content

[Rafal-Chrzanowski-IBM]

Greetings, sorry for the delayed response.

Generational models, like meta-llama/llama-3-1-8b-instruct, do not support function embedding.
You can find a list of currently supported embedding models using commands mentioned in the documentation:
https://ibm.github.io/watsonx-ai-python-sdk/fm_embeddings.html#EmbeddingModels