Nathan Sarrazin
Nathan Sarrazin
 Answers can have an error now and it gets displayed in the front-end when they do. This will hopefully make debugging easier
If you ask an answer and switch conversations while it's answering you, the answer will follow you around and continue streaming. If you refresh the page afterwards, it's gone and...
Currently it's a bit of a mess with little to no structure. I'll be working on making things a bit more structured and expendable.
Title describes it. It probably has something to do with the client closing the connection and therefore the API server has an issue and doesn´t save the output to the...
Pretty low hanging fruit with the wrapper we have, would be great to create a [custom LangChain LLM](https://langchain.readthedocs.io/en/latest/modules/llms/examples/custom_llm.html) wrapper for `llama.cpp`. Then we could use it in the API and...
Would be cool to have in the API.
This PR makes use of the `llama-cpp-python` python bindings. It comes with a custom langchain LLM wrapper for `llama-cpp-python` that supports streaming. The MongoDB requirement is dropped, should improve image...
Big PR. Functionally: * You can now keep conversations going without reloading the entire prompt. * You can have multiple conversations answering at once. * Each conversation has its own...
## Why Since the model has a knowledge cutoff and has a tendency to hallucinate it would be great to complement the chat history with relevant results from the web....
 It's hard to distinguish between past chats and the menu items when the bar is full. I think a low hanging fruit would be to add: * A horizontal...