v0.0.1
A simple console app using llama3-8b-8192 via Express server & Groq.com library.
Open a terminal and clone this repo:
Nav into app directory:
- cd jane-template
Set up the node server:
- cd /res/transformer
- Configure .env with "PORT" and "GROQ_API" key
- use "npm install" to install modules and depenencies
- use "npm run start" start the node server
Compile the executable:
In a new terminal, nav to the app's root directory: "/jane-template"
Compile the app file with:
- gcc init.c -o app -lcurl -lcjson
Make the compiled app file executeable:
- run "chmod +x app"
Initialize the chat
- run "./app" to begin speaking with your new chatbot directly in the terminal
init.c & application binary:
- The main file to initialize the app, handle input/output, and manage conversation context
- Forwards input from terminal to transformer via node.js server and prints response
- You may edit init.c to change the prompt or print speed to your liking
- If compilation fails initially, check init.c to ensure you've installed the required dependencies
- or run "sudo apt-get install libcurl4-openssl-dev libjson-c-dev" to install them (on debian/ubuntu)
README.md:
- (You are here)
/res:
- /res/transformer
- Simple Node/Express server for handling the Groq library integration
- Edit server file to adjust token use and other Groq library options
- See package.json for dependency and version details
NEXT: v0.0.2, Improved context/memory and character/knowledgebank
This project is licensed under the MIT License. Please see the LICENSE file in the repository for details on third-party libraries and their respective licenses.
Built with: