from llama_index.llms import LangChainLLM llm = LangChainLLM(<llm from langchain>) # pretty sure cohere has a 2048 context windo service_context = ServiceContext.from_defaults(llm=llm, context_window=2048)
