/Yolov5-FastApI-Web

This is a FastAPI web application that allows a user to upload images to perform inference using a pretrained YOLOv5 model, and receive results in JSON format.

Primary LanguageHTMLMIT LicenseMIT

YOLOv5 FastAPI Web

This repository hosts a FastAPI web application allowing users to upload images for inference using a pre-trained YOLOv5 model, receiving results in JSON format.

YOLOv5 FastAPI Web

Installation

Ensure Python 3.10 or later is installed, along with dependencies listed in requirements.txt, including torch>=1.7. You can install dependencies using:

pip install -r requirements.txt

Example Usage

Minimal FastAPI Example

Navigate to the client_server_example folder for a minimal client/server wrapper of YOLOv5 with FastAPI and HTML forms.

  1. Run the server with:

    python server_minimal.py

    or

    uvicorn server_minimal:app --reload

  2. Test the server with:

    python client_minimal.py

    or by navigating to localhost:8000 in your web browser and selecting "Try It Out" in the POST request section.

Inference Methods

  1. Initialize the server with:

    python server.py

    (use --help for other arguments) or

    uvicorn server:app --reload

  2. Test the server:

    • Use client.py to upload images and receive JSON inference results.
    • Open localhost:8000 in your web browser to use the web form for image upload and model selection.
    • Open http://localhost:8000/drag_and_drop_detect for a drag and drop interface.

Models are automatically downloaded on first use and cached on disk.

City Street Results

API Documentation

Auto-generated API documentation is available at localhost:8000/docs. The API route provides JSON inference results.

Developer Notes

server.py

Contains FastAPI server code and helper functions.

Jinja2 Frontend Templates (/templates folder)

File Description
layout.html Base template with navbar common to all pages. home.html and drag_and_drop_detect.html extend this template.
home.html Basic web form for uploading images, model selection, and inference. Server renders bbox image and returns results via templates/show_results.html.
drag_and_drop_detect.html Implements a Drag & Drop interface for image upload. Sends image and parameters to server's /detect endpoint, rendering image + bboxes in the browser. Box labels are raised above the outline to prevent overlap.

Credits

This repository is a wrapper around YOLOv5 from Ultralytics: GitHub - ultralytics/yolov5

Results_to_json function is modified from: GitHub Gist