yolov5-fastapi

This machine learning model will perform object detections by allowing a user to upload images and get back results in JSON or image format.

Repo Breakdown

.
├── Dockerfile # For containerization of this application
├── README.md
├── main.py # The entry point of the program
├── model/ # The model files
│   └── best.pt # The trained custom model generated from YOLOv5
├── requirements.in # To generate requirements.txt
├── requirements.txt # Usage: pip install -r requirements.txt
├── segmentation.py # importing local YOLOv5 and scale image
└── yolov5 # YOLOv5 repo from https://github.com/ultralytics/yolov5

Prerequisites

Install the packages required for this project

pip install -r requirements.txt

Teck Stack

• Tech-stack
  • FastAPI - Fast API framework for serving authentication API with the asynchronous pattern.
  • YOLOv5 - For training custom object detection models
  • Docker -Building container images.

To start for this project

uvicorn main:app --reload --host 0.0.0.0 --port 8000

To build the docker image

docker build -t yolov5-fastapi .

To run the docker container of the image

docker run -it --rm -p 8000:8000 yolov5-fastapi

The application will be available on http://localhost:8000/

Todo:

• Add dockerfile
• Add authentication to the api
• Add feature: store detected result into a database
• Add more tests