Stable diffusion API 📚🔎

Local installation

To prepare the project, the following steps will be necessary:

Install git locally;

On your local server, run a git clone command;

Local execution

To execute the project, there's two ways to do it:

First download python 3. After, install the dependencies by typing pip install -r requirements.txt and run python main.py on your terminal if you just want to verify all the routes and see if the API algorithm is working;

If you already have Docker installed on you machine, run docker-compose -f docker-compose.dev.yml up --build -d on the terminal to create an image and container for the aplication (and for the cache server when it's available). This way allows you to install all application dependencies as well as automate the API testing process through docker;

In both flows, after uploading the application you can access it at localhost:8000/docs, as well as check the logs informations at localhost:8000/stable-diffusion/logs.

Feature

The repository features the following tool:

This repository contains a tool in REST API format to generate images using the public stable diffusion model trained by the hugging face company 🤗.

Routes

AI image generator

POST /stable-diffusion/image
Request body

{
"text": "string"
}

Media type: application/json
Response

{
"image": '_filename'.jpg
}

Metrics

GET /stable-diffusion/metrics
Media type: text/plain
Response

"string"

Logs

GET /stable-diffusion/logs
Media type: text/plain
Response

"string"

Health

GET /stable-diffusion/health
Media type: application/json
Response

{
"health": "string"
}

Example of use:

After asking for a specific word, the API will return a AI created image related to the respective requested word, as shown in the following examples:

Image

POST http://localhost:8000/stable-diffusion/image

Request Body:

  {
  "text": "A realistic image of Synthwave futuristic city"
  }

Response Body: