• 2021.3.21: 对模型结构进行细微调整优化，更新Yolo-Fastest-1.1模型
• 2021.3.19: NCNN Camera Demo https://github.com/dog-qiuqiu/Yolo-Fastest/tree/master/sample/ncnn
• 2021.3.16: 修复分组卷积在某些旧架构GPU推理耗时异常的问题

⚡Yolo-Fastest⚡

• Simple, fast, compact, easy to transplant
• A real-time target detection algorithm for all platforms
• The fastest and smallest known universal target detection algorithm based on yolo
• Optimized design for ARM mobile terminal, optimized to support NCNN reasoning framework
• Based on NCNN deployed on RK3399 ,Raspberry Pi 4b... and other embedded devices to achieve full real-time 30fps+

• 中文介绍https://zhuanlan.zhihu.com/p/234506503
• 相比AlexeyAB/darknet，此版本的darknet修复分组卷积在某些旧架构GPU推理耗时异常的问题(例如1050ti:40ms->4ms速度提升10倍)，强烈建议用此仓库框架训练模型
• Compared with AlexeyAB/darknet, this version of darknet fixes the problem of abnormal time-consuming inference of grouped convolution in some old architecture GPUs (for example, 1050ti:40ms->4ms speed up 10 times), it is strongly recommended to use this warehouse framework for training model
• Darknet CPU推理效率优化不好，不建议使用Darknet作为CPU端的推理框架，建议使用NCNN
• Darknet CPU reasoning efficiency optimization is not good, it is not recommended to use Darknet as the CPU side reasoning framework, it is recommended to use ncnn
• Based on pytorch training framework: https://github.com/dog-qiuqiu/yolov3

Evaluating indicator/Benchmark

Network	COCO mAP(0.5)	Resolution	Run Time(Ncnn 4xCore)	Run Time(Ncnn 1xCore)	FLOPS	Params	Weight size
Yolo-Fastest-1.1	24.40 %	320X320	5.59 ms	7.52 ms	0.252BFlops	0.35M	1.4M
Yolo-Fastest-1.1-xl	34.33 %	320X320	9.27ms	15.72ms	0.725BFlops	0.925M	3.7M
Yolov3-Tiny-Prn	33.1%	416X416	%ms	%ms	3.5BFlops	4.7M	18.8M
Yolov4-Tiny	40.2%	416X416	23.67ms	40.14ms	6.9 BFlops	5.77M	23.1M

• Test platform Mi 11 Snapdragon 888 CPU，Based on NCNN
• COCO 2017 Val mAP（no group label）
• Suitable for hardware with extremely tight computing resources
• This model is recommended to do some simple single object detection suitable for simple application scenarios

Yolo-Fastest-1.1 Multi-platform benchmark

Equipment	Computing backend	System	Framework	Run time
Mi 11	Snapdragon 888	Android(arm64)	ncnn	5.59ms
Mate 30	Kirin 990	Android(arm64)	ncnn	6.12ms
Meizu 16	Snapdragon 845	Android(arm64)	ncnn	7.72ms
Development board	Snapdragon 835(Monkey version)	Android(arm64)	ncnn	20.52ms
Development board	RK3399	Linux(arm64)	ncnn	35.04ms
Raspberrypi 3B	4xCortex-A53	Linux(arm64)	ncnn	62.31ms
Orangepi Zero Lts	H2+ 4xCortex-A7	Linux(armv7)	ncnn	550ms
Nvidia	Gtx 1050ti	Ubuntu(x64)	darknet	4.73ms
Intel	i7-8700	Ubuntu(x64)	ncnn	5.78ms

• The above is a multi-core test benchmark
• The above speed benchmark is tested by big core in big.little CPU
• Raspberrypi 3B enable bf16s optimization，Raspberrypi 64 Bit OS
• Rk3399 needs to lock the cpu to the highest frequency, ncnn and enable bf16s optimization

Pascal VOC performance index comparison

Network	Model Size	mAP(VOC 2007)	FLOPS
Tiny YOLOv2	60.5MB	57.1%	6.97BFlops
Tiny YOLOv3	33.4MB	58.4%	5.52BFlops
YOLO Nano	4.0MB	69.1%	4.51Bflops
MobileNetv2-SSD-Lite	13.8MB	68.6%	&Bflops
MobileNetV2-YOLOv3	11.52MB	70.20%	2.02Bflos
Pelee-SSD	21.68MB	70.09%	2.40Bflos
Yolo Fastest	1.3MB	61.02%	0.23Bflops
Yolo Fastest-XL	3.5MB	69.43%	0.70Bflops
MobileNetv2-Yolo-Lite	8.0MB	73.26%	1.80Bflops

• Performance indicators reference from the papers and public indicators in the github project
• MobileNetv2-Yolo-Lite: https://github.com/dog-qiuqiu/MobileNet-Yolo#mobilenetv2-yolov3-litenano-darknet

Yolo-Fastest-1.1 Pedestrian detection

Equipment	System	Framework	Run time
Raspberrypi 3B	Linux(arm64)	ncnn	62ms

• Simple real-time pedestrian detection model based on yolo-fastest-1.1
• Enable bf16s optimization，Raspberrypi 64 Bit OS

Demo

Compile

How to compile on Linux

• This repo is based on Darknet project so the instructions for compiling the project are same (https://github.com/MuhammadAsadJaved/darknet#how-to-compile-on-windows-legacy-way)

Just do make in the Yolo-Fastest-master directory. Before make, you can set such options in the Makefile: link

• GPU=1 to build with CUDA to accelerate by using GPU (CUDA should be in /usr/local/cuda)
• CUDNN=1 to build with cuDNN v5-v7 to accelerate training by using GPU (cuDNN should be in /usr/local/cudnn)
• CUDNN_HALF=1 to build for Tensor Cores (on Titan V / Tesla V100 / DGX-2 and later) speedup Detection 3x, Training 2x
• OPENCV=1 to build with OpenCV 4.x/3.x/2.4.x - allows to detect on video files and video streams from network cameras or web-cams
• Set the other options in the Makefile according to your need.

Test/Demo

*Run Yolo-Fastest , Yolo-Fastest-xl , Yolov3 or Yolov4 on image or video inputs

Demo on image input

*Note: change .data , .cfg , .weights and input image file in image_yolov3.sh for Yolo-Fastest-x1, Yolov3 and Yolov4

  sh image_yolov3.sh

Demo on video input

*Note: Use any input video and place in the data folder or use 0 in the video_yolov3.sh for webcam

*Note: change .data , .cfg , .weights and input video file in video_yolov3.sh for Yolo-Fastest-x1, Yolov3 and Yolov4

  sh video_yolov3.sh

Yolo-Fastest Test

Yolo-Fastest-xl Test

How to Train

Generate a pre-trained model for the initialization of the model backbone

  ./darknet partial yolo-fastest.cfg yolo-fastest.weights yolo-fastest.conv.109 109

Train

• 交流qq群:1062122604
• https://github.com/AlexeyAB/darknet

  ./darknet detector train voc.data yolo-fastest.cfg yolo-fastest.conv.109 

Deploy

NCNN

NCNN Conversion Tutorial

• Benchmark:https://github.com/Tencent/ncnn/tree/master/benchmark
• NCNN supports direct conversion of darknet models
• darknet2ncnn: https://github.com/Tencent/ncnn/tree/master/tools/darknet

NCNN Sample

• CamSample:https://github.com/dog-qiuqiu/Yolo-Fastest/tree/master/sample/ncnn
• AndroidSample: https://github.com/WZTENG/YOLOv5_NCNN

MNN&TNN&MNN

• https://github.com/dog-qiuqiu/MobileNet-Yolo#darknet2caffe-tutorial
• Based on MNN: https://github.com/geekzhu001/Yolo-Fastest-MNN Run on : raspberry pi 4B 2G Input size : 320320 Average inference time : 0.035s*

ONNX&TensorRT

• https://github.com/CaoWGG/TensorRT-YOLOv4
• It is not efficient to run on Psacal and earlier GPU architectures. It is not recommended to deploy on such devices such as jeston nano(17ms/img), Tx1, Tx2, but there is no such problem in Turing GPU, such as jetson-Xavier-NX Can run efficiently

OpenCV DNN

• https://blog.csdn.net/nihate/article/details/108670542

Thanks

• https://github.com/AlexeyAB/darknet
• https://github.com/Tencent/ncnn

Cite as

dog-qiuqiu. (2021, July 24). dog-qiuqiu/Yolo-Fastest: yolo-fastest-v1.1.0 (Version v.1.1.0). Zenodo. http://doi.org/10.5281/zenodo.5131532