billybsig / demo-0424

Demo

About

This repository contains an implementation of the Yolo object detection model's inference. The endpoint requires uploading an image or video. The code will automatically detect our device (CPU/CUDA). If our device is using CPU, the code will utilize the ONNX model (.onnx), and if our device is using GPU, it will employ the PyTorch model (.pt).

Installation

• clone the repository

    git clone https://github.com/BillyBSig/Demo-0424.git
  

• download the YOLO model following these link

  • pytorch model

      https://drive.google.com/file/d/1FKzxILrtQlQ-wLK_AmxvhjtKl-Yhmakc/view?usp=sharing
    

  • onnx model

      https://drive.google.com/file/d/1LpBmCS7WUSdrRZFHkiFfgmmsujzQInNO/view?usp=sharing
    

• Put these model in model directory

    ├── Lmodel/
    |   ├── yolov8n.onnx
    |   ├── yolov8n.pt
  

• run the code on CLI

    uvicorn main:app --host 0.0.0.0 --port 7373 --reload
  

• The end point will serve locally in http://127.0.0.1:7373/upload

Docker instruction

• build the docker image

    docker build -t yolo_api .
  

• run the docker following this step:

  • Create and/or copy the output directory that contain the video or image output process e.g:

    D:\YoloDemo\detect

  • if we use the GPU cuda, run

      docker run --gpus all -d  -p 7373:7373 -v local//output-detect:/app/detect --name yolo_api_serve yolo_api
    

    • change the local//output-detect with our local output directory e.g :

      docker run --gpus all -d -p 7373:7373 -v D:\YoloDemo\detect:/app/detect --name yolo_api_serve yolo_api

    • the output will using the internal docker path e.g:

        {"path": "/app/detect/2024-04-21-153032/meeting.jpg"}
      

    Then in our local device, output will save in local output directory/2024-04-21-153032/meeting.jpg

  • if we use the CPU only, then run

      docker run -d  -p 7373:7373 -v loca//output-detect:/app/detect --name yolo_api_serve yolo_api
    

    • change the local//output-detect with our local output directory

      e.g :

      docker run -d -p 7373:7373 -v D:\YoloDemo\detect:/app/detect --name yolo_api_serve yolo_api