Can anyone tell me what’s wrong with this code?

def get_answer_from_text(document_id: str, question: str, conversation_history: List[dict]) -> str:
    try:
        Settings.embed_model = HuggingFaceEmbedding(
            model_name="BAAI/bge-small-en-v1.5"
        )
        
        reader = SimpleDirectoryReader(input_files=[os.path.join(UPLOAD_DIR, f"{document_id}.txt")])
        documents = reader.load_data()

        # Combine conversation history with the new question
        combined_text = "n".join([f"Q: {entry['question']}nA: {entry['answer']}" for entry in conversation_history])
        combined_text += f"nQ: {question}"

        llm = HuggingFaceInferenceAPI(model_name = "mistralai/Mistral-7B-Instruct-v0.3", token=access_token)

        Settings.llm = llm

        index = VectorStoreIndex.from_documents(
            documents, 
            embed_model=embed_model, llm=llm 
        )
        
        query_engine = index.as_query_engine(llm = llm)
        answer = query_engine.query(combined_text, top_k=1)
        print(answer)

I have tried using the HuggingFaceEmbedding model as well as HuggingFaceInferenceAPI model but it is still not working with either of the way. I am getting the error: “embed_model” is not defined. Can someone please check this code and help me with this?