yolov5 tensorrt

这个感觉靠谱，还没试：

tensorrtx/yolov5 at master · wang-xinyu/tensorrtx · GitHub

这个不错，c++调通的版本：

yolov5转tensorrt c++_jacke121的专栏-CSDN博客

python tensorrt:

激活函数：hard_sigmoid

https://github.com/TrojanXu/yolov5-tensorrt

https://github.com/wang-xinyu/tensorrtx

The Pytorch implementation is ultralytics/yolov5.

Currently, we support yolov5 v1.0(yolov5s only), v2.0, v3.0 and v3.1.

For yolov5 v3.1, please visit yolov5 release v3.1, and use the latest commit of this repo.
For yolov5 v3.0, please visit yolov5 release v3.0, and use the latest commit of this repo.
For yolov5 v2.0, please visit yolov5 release v2.0, and checkout commit '7cd092d' of this repo.
For yolov5 v1.0, please visit yolov5 release v1.0, and checkout commit '0504551' of this repo.

Config

Choose the model s/m/l/x by NET macro in yolov5.cpp
Input shape defined in yololayer.h
Number of classes defined in yololayer.h, DO NOT FORGET TO ADAPT THIS, If using your own model
FP16/FP32 can be selected by the macro in yolov5.cpp
GPU id can be selected by the macro in yolov5.cpp
NMS thresh in yolov5.cpp
BBox confidence thresh in yolov5.cpp
Batch size in yolov5.cpp

v2也是leakyrelu。

2.0:这个是匹配3.0的版本，用的leakyrelu,可以检测，v3.0自己训练的精度比较低

GitHub - BaofengZan/yolov5_2.0-TensorRt: U版yolov5 2.0的tensorrt加速

https://github.com/AIpakchoi/yolov5_tensorrt/blob/110762eea4a7a53a91bbce35f94239136db157d4/yolov5l/common.hpp

GitHub - baituhuangyu/yolov5-tensorrt: yolov5 tensorrt inference

https://github.com/Thinker-or-Dreamer/UAV-And-RobotArm/tree/master/yolov5

HardSwishLayer_TRT

https://github.com/hlld/tensorrt-yolov5

linux的：

2020.10.23 17天以前更新的，激活函数：kHARD_SIGMOID

https://github.com/wang-xinyu/tensorrtx/tree/master/yolov5

yololayer.h

修改自己类别个数：

static constexpr int CLASS_NUM = 1;

下面是阈值参数，nms阈值参数：

#define USE_FP16  // comment out this if want to use FP32
#define DEVICE 0  // GPU id
#define NMS_THRESH 0.4
#define CONF_THRESH 0.3std::cerr << "yolov5_rt.exe -s s // serialize model to plan file" << std::endl;
std::cerr << "yolov5_rt.exe -e s -c 0 // detect cam" << std::endl;
std::cerr << "yolov5_rt.exe -e s -d samples  // deserialize plan file and run inference" << std::endl;

后面3行是调用demo，分两步，编译和执行

问题：原版网络检测出来的框没问题

自己训练的，tensorrt检测的与pytorch检测出来的有偏差，原因还未找到。

可能是anchors的原因，但是没找到证据。

yolov5 tensorrt相关推荐

YOLOv5+TensorRT+Win11(Python版)
快速上手YOLOv5 快速上手YOLOv5 一.YOLOv5算法 1. 算法对比 (1)传统目标检测方法 (2)基于深度学习的目标检测算法 (2-1)Two-Stage(R-CNN/Fast R-CN ...
win10 yolov5 tensorRT 部署
参考: 1.Win10-YOLOv5实战+TensorRT部署+VS2019编译(小白教程~易懂易上手)-超详细] 2.YOLOV5(Pytorch)目标检测实战:TensorRT加速部署视频目录 ...
Yolov5 Tensorrt win10 C++ 部署2022
所有前置环境包都在Q群公告中的网盘链接中 YOLO交流Q群 936625049 目录 1.前置条件 2.VS 2017下载及配置 3.CUDA部分 4.生成wts模型 5.Opencv配置 6.Ten ...
yolov5 tensorrt 精度对齐总结
本文对c++推理的yolov5 v6.1代码进行精度对齐实现,以yolov5-l为例. yolov5:https://github.com/ultralytics/yolov5 tensorrtx:G ...
yolov5 tensorrt加速
视频教程 https://www.bilibili.com/video/BV113411J7nk?p=1 Github仓库地址 https://github.com/Monday-Leo/Yolov5 ...
NVIDIA Jetson YOLOv5 tensorRT部署和加速 C++版
前言在实现NVIDIA Jetson AGX Xavier 部署YOLOv5的深度学习环境,然后能正常推理跑模型后:发现模型速度不够快,于是使用tensorRT部署,加速模型,本文介绍C++版本的. ...
YOLOV5 v6.1更新 | TensorRT+TPU+OpenVINO+TFJS+TFLite等平台一键导出和部署
点击上方"3D视觉工坊",选择"星标" 干货第一时间送达作者丨ChaucerG 来源丨集智书童 yolov5 release 6.1版本增加了TensorRT ...
tensorrt部署YOLOv5模型记录【附代码，支持视频检测】
训练出来的模型最终都需要进行工业部署,现今部署方案有很多,tensorflow和pytorch官方也都有发布,比如现在pytorch有自己的Libtorch进行部署[可以看我另一篇文章有讲利用Libt ...
TX2安装pytorch+TensorRT+yolov5实现实时检测
已完成的环境配置: TX2刷机后,完成了opencv4.5.1的编译:Ubuntu18.04安装opencv4.5.1+contrib 支持cuda加速(附带编译好的opencv4.5.1及缺失文件) ...

yolov5 tensorrt

Config

yolov5 tensorrt相关推荐

最新文章

热门文章