Skip to content

Latest commit

 

History

History
51 lines (43 loc) · 2.42 KB

README.md

File metadata and controls

51 lines (43 loc) · 2.42 KB
Raw

Reflection 70B Chat

Reflection 70B Chat UI Image

Description

  • This GitHub repository integrates Reflection Llama-3.1 70B, the world's top open-source large language model (LLM), in a Streamlit-based user interface. The LLM is trained using a new technique called Reflection-Tuning, which enables it to detect and correct mistakes in its own reasoning.

NOTE!!

  • As the API providers for this model (Reflection-70B) have removed the free version, this implementation will not work. But feel free to look into the implementation of the code.

Libraries Used

  • requests
  • streamlit
  • langchain
  • python-dotenv
  • langchain_community

File and Folder Explanation

  1. lib: contains configuration and utility code files.
  2. images: contains UI image.
  3. app.py: main function that will run the Streamlit UI.
  4. lib/conifg.py: contains functions to load our environment variables and get our api keys.
  5. lib/utils.py: contains a function to setup our llm, accept user query and provide response.

Installation

  1. Prerequisites
    • Git
    • Command line familiarity
  2. Clone the Repository: git clone https://github.com/NebeyouMusie/Reflection-70B-Chat.git
  3. Create and Activate Virtual Environment (Recommended)
    • python -m venv venv
    • source venv/bin/activate for Mac and venv/bin/activate for Windows
  4. Navigate to the projects directory cd ./Reflection-70B-Chat using your terminal
  5. Install Libraries: pip install -r requirements.txt
  6. Enter your OPENROUTER_API_KEY in the example.env file then change the file to .env. You can get your OPENROUTER_API_KEY from here.
  7. run streamlit run app.py
  8. open the link that appears on your terminal in your preferred browser.

Usage

  • Start by typing your question in the Chat Input located at the bottom of the app.
  • The LLM will provide a response, which will include the action it took to give the answer to the user along with the main response(output).
  • Note that it won't remember previous user interactions (NO MEMORY) as that functionality has not been added.

Collaboration

  • Collaborations are welcomed ❤️

Acknowledgments

Contact