Skip to content

gajojr/gpt-steroid

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

55 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Chat GPT Steroid

Chat GPT Steroid is an enhanced version of the Chat GPT language model built on top of OpenAI API. It allows up to 32K tokens for both input and output, which is 4 times larger than the Chat GPT 4. The data is stored locally in the browser's IndexedDB, and it allows for file uploads to fine-tune the model. The project also has a local setup, making it easy to run the application on your local machine.

Features (built on top of chatGPT 4)

  • File uploads for fine-tuning
  • Create fine tune file via gui
  • 32K token input and output
  • Local setup
  • IndexedDB data storage
  • Connection to Auto GPT for automatic prompting

Getting Started

  1. Install Node.js
  2. Install Docker
  3. Clone the repository from GitHub: git clone https://github.com/gajojr/gpt-steroid.git
  4. Navigate to the cloned repository
  5. Clone the AutoGPT repo from root: git clone https://github.com/Significant-Gravitas/Auto-GPT.git
  6. Create an OpenAI API key with a paid account
  7. Navigate to server directory, create a .env file and set OPENAI_API_KEY=YOUR_API_KEY, where YOUR_API_KEY is the value of the API key from OpenAI
  8. Navigate to frontend directory, create a .env file and set REACT_APP_SERVER_URL=http://localhost:8000
  9. Navigate to Auto-GPT directory and create .env file and set OPENAI_API_KEY=YOUR_API_KEY
  10. Run npm install in root
  11. Run npm install in frontend and server to install dependencies
  12. Start docker
  13. Run npm run dev from root to start the application

Additional Features (in progress)

  • Access to the internet

Contributing

We welcome contributions from anyone! Please feel free to submit a pull request or open an issue on GitHub.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published