Installing YOLOv4 with Anaconda

This project uses the YOLOv4/Anaconda Setup by The AI Guy. This is part one of this series:

• Installing YOLOv4 using Anaconda
• License Plate Recognition with YOLOv4, OpenCV and Tesseract
• Crop License Plate from the Original Image

• Yolo4 Object Detection Setup
  • Dependencies
    • Anaconda
    • Git
    • Source Repository
    • Pre-trained Weights
  • Virtual Environment
    • CPU
    • GPU
  • Convert weights to TensorFlow Format
• Run YOLOv4
  • Run YOLOv4 Tensorflow Model on an Image
  • Run YOLOv4 from a Video File
  • Run YOLOv4 on a Webcam Stream
  • Run YOLOv4 on a INSTAR IP Camera Stream

Yolo4 Object Detection Setup

Dependencies

Anaconda

Install Conda:

wget https://repo.anaconda.com/archive/Anaconda3-2021.11-Linux-x86_64.sh
chmod +x Anaconda3-2021.11-Linux-x86_64.sh
bash Anaconda3-2021.11-Linux-x86_64.sh

Add conda to your PATH variables (~/.bashrc, ~/.zshrc, etc) - don' t forget to source it afterwards:

export PATH="/home/myuser/anaconda3/bin:$PATH"

Verify:

conda --version
conda 4.10.3

Initialize your shell with:

conda init <SHELL_NAME>

Currently supported shells are:

• bash
• fish
• tcsh
• xonsh
• zsh
• powershell

Git

conda install -c anaconda git

Source Repository

And clone the source code from Github:

git clone https://github.com/mpolinowski/yolov4-custom-functions.git

Pre-trained Weights

YOLOv4 comes pre-trained and able to detect 80 classes. For easy demo purposes we will use the pre-trained weights. Download pre-trained yolov4.weights file:

• yolov4.weights

Copy and paste yolov4.weights from your downloads folder into the 'data' folder of the repository.

Virtual Environment

Enter the repository and either run the CPU or GPU setup - the latter requires an NVIDIA graphic card with CUDA support:

CPU

conda env create -f conda-cpu.yml
conda activate yolov4-cpu

GPU

conda env create -f conda-gpu.yml
conda activate yolov4-gpu

Convert weights to TensorFlow Format

To implement YOLOv4 using TensorFlow, first we convert the .weights into the corresponding TensorFlow model files and then run the model.

Convert darknet weights to tensorflow:

python save_model.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4-416 --input_size 416 --model yolov4

Run YOLOv4

Run YOLOv4 Tensorflow Model on an Image

python detect.py --weights ./checkpoints/yolov4-416 --size 416 --model yolov4 --images ./data/images/Old_Town.jpg

Run YOLOv4 from a Video File

python detect_video.py --weights ./checkpoints/yolov4-416 --size 416 --model yolov4 --video ./data/video/cars.mp4 --output ./detections/results.avi

Run YOLOv4 on a Webcam Stream

Connect your webcam and run YOLOv4:

python detect_video.py --weights ./checkpoints/yolov4-416 --size 416 --model yolov4 --video 0 --output ./detections/results.avi

Run YOLOv4 on a INSTAR IP Camera Stream

python detect_video.py --weights ./checkpoints/yolov4-416 --size 416 --model yolov4 --video http://192.168.0.80:80/mjpegstream.cgi?-chn=11&-usr=admin&-pwd=instar