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.
- 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
- Install Node.js
- Install Docker
- Clone the repository from GitHub:
git clone https://github.com/gajojr/gpt-steroid.git
- Navigate to the cloned repository
- Clone the AutoGPT repo from root:
git clone https://github.com/Significant-Gravitas/Auto-GPT.git
- Create an OpenAI API key with a paid account
- Navigate to server directory, create a
.env
file and setOPENAI_API_KEY=YOUR_API_KEY
, whereYOUR_API_KEY
is the value of the API key from OpenAI - Navigate to frontend directory, create a
.env
file and setREACT_APP_SERVER_URL
=http://localhost:8000 - Navigate to Auto-GPT directory and create
.env
file and setOPENAI_API_KEY
=YOUR_API_KEY
- Run
npm install
in root - Run
npm install
in frontend and server to install dependencies - Start docker
- Run
npm run dev
from root to start the application
- Access to the internet
We welcome contributions from anyone! Please feel free to submit a pull request or open an issue on GitHub.