Home

Darknet (neural network framework)

YOLOv4 model zoo: https://github.com/AlexeyAB/darknet/wiki/YOLOv4-model-zoo

FAQ - frequently asked questions: FAQ - frequently asked questions

Basic terms:

• cfg-files - are structures of neural networks: https://github.com/AlexeyAB/darknet/tree/master/cfg
• weights-files - are weights for correspond cfg-file, can be downloaded from: https://pjreddie.com/darknet/

There are several tasks that are implemented in the Darknet framework:

• Yolo detection: https://pjreddie.com/darknet/yolo/
• ImageNet classification: https://pjreddie.com/darknet/imagenet/
• Text Generation: https://pjreddie.com/darknet/rnns-in-darknet/
• GO playing: https://pjreddie.com/darknet/darkgo-go-in-darknet/
• GAN Nightmare: https://pjreddie.com/darknet/nightmare/

Training and Evaluation of speed and accuracy on MS COCO

1. Train Classifier on ImageNet (ILSVRC2012): https://github.com/AlexeyAB/darknet/wiki/Train-Classifier-on-ImageNet-(ILSVRC2012)

2. Train YOLOv4 Detector on MS COCO (trainvalno5k 2014) dataset: https://github.com/AlexeyAB/darknet/wiki/Train-Detector-on-MS-COCO-(trainvalno5k-2014)-dataset

3. How to evaluate accuracy and speed of YOLOv4 on MS COCO dataset: https://github.com/AlexeyAB/darknet/wiki/How-to-evaluate-accuracy-and-speed-of-YOLOv4

Training and Evaluation on other Datasets

1. Train and Evaluate Detector on Pascal VOC (VOCtrainval 2007 2012) dataset: https://github.com/AlexeyAB/darknet/wiki/Train-and-Evaluate-Detector-on-Pascal-VOC-(VOCtrainval-2007-2012)-dataset

Meaning of configuration parameters in the cfg-files

1. CFG-Parameters in the [net] section: https://github.com/AlexeyAB/darknet/wiki/CFG-Parameters-in-the-%5Bnet%5D-section

2. CFG-Parameters in the different layers: https://github.com/AlexeyAB/darknet/wiki/CFG-Parameters-in-the-different-layers

Use Yolo v3 in other frameworks

CLICK ME - Use Yolo in other frameworks

• OpenCV-dnn the fastest implementation for CPU (x86/ARM-Android), OpenCV can be compiled with OpenVINO-backend for running on (Myriad X / USB Neural Compute Stick / Arria FPGA), use yolov3.weights/cfg with: C++ example or Python example

• Converting Yolo v3 models to TensorFlow and OpenVINO(IR) models: https://github.com/AlexeyAB/darknet/wiki/Converting-Yolo-v3-models-to-TensorFlow-and-OpenVINO(IR)-models

  • TensorFlow: convert yolov3.weights/cfg files to yolov3.ckpt/pb/meta: by using mystic123 or jinyu121 projects, and TensorFlow-lite
  • Intel OpenVINO 2019 R1: (Myriad X / USB Neural Compute Stick / Arria FPGA): read this manual

• PyTorch > ONNX > CoreML > iOS how to convert cfg/weights-files to pt-file: ultralytics/yolov3 and iOS App

• TensorRT for YOLOv3 (-70% faster inference): Yolo is natively supported in DeepStream 4.0 read PDF

• TVM - compilation of deep learning models (Keras, MXNet, PyTorch, Tensorflow, CoreML, DarkNet) into minimum deployable modules on diverse hardware backends (CPUs, GPUs, FPGA, and specialized accelerators): https://tvm.ai/about

• OpenDataCam - It detects, tracks and counts moving objects by using Yolo: https://github.com/opendatacam/opendatacam#-hardware-pre-requisite

• Netron - Visualizer for neural networks: https://github.com/lutzroeder/netron