Is there API to limit the number number of tokens in request

[ydsgroup]

We are trying to fetch the context and answers from LLM as below

    llm = ChatOpenAI(temperature = 0.0)
    retriever = db.as_retriever()
    qa_stuff = RetrievalQA.from_chain_type(
    llm=llm,chain_type="stuff",retriever=retriever,verbose=True)
    result = qa_stuff.run(user_input)

Some time this api retrieve large context which exceed the number of tokens .
I am using the in memory docarray db for context search

I know we can limit the token by manual running the query resultant documents and iterating over documents and combine them and before combine we can check length of tokens .

Is there any way I can limit the by some chain apis . also can how can I switch over strategy of chain_type on the go as different document need different strategies .

Langchain is mind blowing framework simplifying the LLM work . Thanks lot for such framework .
I would really appreciate for any pointers to my questions .

[adriens]

That's a very good idea 💡

[ydsgroup]

I got couple of answers at new session

https://learn.deeplearning.ai/langchain-chat-with-your-data

Thanks lot for educating people with very good session .