Bacground Removal and Product Placement API

Description

This API is used to remove the background of an image and place the product on a desired background. Background generation is done using Stable Diffusion models (https://huggingface.co/stabilityai/stable-diffusion-2-inpainting). The API is built using FastAPI and deployed using Docker.

Requirements

Docker
Nvidia GPU (optional but recommended)

Usage

1. Clone the repository

git clone https://github.com/arda-num/stable-diffusion.git

2.

(OPTION 1) Build the docker image with docker-compose

docker-compose up

(OPTION 2) Install the dependencies and run the API

pip install -r requirements.txt
uvicorn main:app --reload

3. Test the API (Sample Requests)

Inpainting

curl -o download.png --location --request POST 'http://127.0.0.1:8000/inpaint' \
--form 'prompt="a bottle of perfume with confetti in the background, a digital rendering by Jeff Koons, cg society contest winner, new objectivity, vray tracing, dynamic composition, rendered in unreal engine"' \
--form 'image=@"path/to/image.png"'

Remove Background

curl -o download.png --location --request POST 'http://127.0.0.1:8000/removebackground' \
--form 'image=@"path/to/image.png"'

FAQ

If you are getting an error called "stable-diffusion-api-1 exited with code 137",

This is because the docker container is running out of memory. You can increase the memory limit of the container by going to Docker Desktop -> Settings -> Resources -> Advanced -> Memory and increasing the memory limit

you can try changing the limits and reservations in docker-compose.yml file.

deploy:
    resources:
    limits:
        memory: 500M
    reservations:
        memory: 128M