This is a Next.js project bootstrapped with create-next-app.
This is an Avatar creator made by building a frontend wrapper on top of a stable diffusion model which was trained with my pictures. The model hosted on huggingface, expects a prompt as an input with 'Aryan' and generates an image. The project was created as one of Buildspace's weekend hack to explore functionality of Stable Diffusion and how developers can build on top of it.
If you want to create your own model, you can use this public notebook to feed the model your own images. (All thanks to Buildspace for the ipynb notebook!)
Head over to my hosted app to see it work.
If you wanna dig deeper and understand how I created the app, go ahead and follow these steps
To run the app locally, you need to use the sd-model, on top of which this wrapper interface has been created. Clone this repository on huggingface. And get your API key to run the model
Create a .env file and store the API Key in this variable.
HF_AUTH_KEY = your_keyRun the development server and see the trained stable diffusion model live in action!
npm run dev
# or
yarn dev