10000 GitHub - Nikalb/darknet: YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
/ darknet Public
forked from AlexeyAB/darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )

License

Notifications You must be signed in to change notification settings

Nikalb/darknet

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Yolov4, Yolo-Fish, Docker container on a Server

Version 1.0, built by J. Giessmann, R. Subiza and N. Albers

The goal was it to set up a version of Yolov4 with the loaded Yolo-Fish model inside a docker container on a linux server to detect zebrafish inside of aquariums. Ultimately counting them, as this data is needed for research. To access the container more easily, a python web server was added, to upload and download data more easily.

Steps for setup

How to download weights

How to change the weights

  • change the file the stated directory, for example fo 73B4 r trained weights
  • change the variable >>weights_path<< in darknet_main.py

Configured original files

  • docker-compose.yml
  • Dockerfile.cpu
  • cocoo.data
  • coco.names
  • Makefile

Using GPU

  • Adapt the Dockerfile.gpu in a similar fashion to Dockerfile.cpu
  • Change the values in Makefile to support GPU, make sure all drivers are and other dependencies are installied on the server
  • Change the docker-compose.yml for yolo-gpu in a similar fashion to yolo-cpu, with the entrypoint, volume and ports.

Ideas for future features

About

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • C 62.8%
  • Cuda 14.9%
  • C++ 12.7%
  • Python 3.4%
  • PowerShell 2.7%
  • CMake 1.7%
  • Other 1.8%
0