# Llama-2- chatbot with FAISS using langchain

**URL:** <https://discuss.huggingface.co/t/llama-2-chatbot-with-faiss-using-langchain/109953>\
**Category:** Beginners\
**Created:** [October 3, 2024, 11:45am UTC](https://discuss.huggingface.co/t/llama-2-chatbot-with-faiss-using-langchain/109953 "2024-10-03T11:45:03Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Manickam-yenmin](https://avatars.discourse-cdn.com/v4/letter/m/898d66/32.png) [@Manickam-yenmin](https://discuss.huggingface.co/u/Manickam-yenmin)\
**Post date:** [October 3, 2024, 11:45am UTC](https://discuss.huggingface.co/t/llama-2-chatbot-with-faiss-using-langchain/109953/1 "2024-10-03T11:45:03Z")

</div>

I’m creating a chatbot using FAISS and the Llama-2-7b-ggml model. In this, I’m encountering follow-up questions from users without introducing the vector-pipeline. Is it possible to handle this directly, or do I need to take a different approach to manage such scenarios?
