ATOM_resnet18_uav数据集 用自己训练的模型进行测试

最新推荐文章于 2023-06-22 08:30:08 发布
最新推荐文章于 2023-06-22 08:30:08 发布
阅读量837 收藏 2
点赞数 1
分类专栏： 笔记 # 目标跟踪 # pytorch 文章标签： 深度学习 pytorch 神经网络
zzz
本文链接：https://blog.csdn.net/weixin_45700507/article/details/108086948
版权
笔记 同时被 3 个专栏收录
32 篇文章 0 订阅
订阅专栏
pytorch
25 篇文章 0 订阅
订阅专栏
目标跟踪
20 篇文章 1 订阅
订阅专栏

run_experiment.py debug=0 atom default range(1)(一个跟踪器) 结果保存在pytracking.tracking_results_me.atom.default_000

Running:  myexperiments  uav_test
Evaluating    1 trackers on   123 sequences
Tracker: atom default 0 ,  Sequence: uav_bike1
/home/zhuzhu/anaconda3/envs/pytracking/lib/python3.7/site-packages/torch/nn/functional.py:2506: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  "See the documentation of nn.Upsample for details.".format(mode))
/home/zhuzhu/anaconda3/envs/pytracking/lib/python3.7/site-packages/setuptools/distutils_patch.py:26: UserWarning: Distutils was imported before Setuptools. This usage is discouraged and may exhibit undesirable behaviors or errors. Please use Setuptools' objects directly or at least import Setuptools first.
  "Distutils was imported before Setuptools. This usage is discouraged "
Using /tmp/torch_extensions as PyTorch extensions root...
Detected CUDA files, patching ldflags
Emitting ninja build file /tmp/torch_extensions/_prroi_pooling/build.ninja...
Building extension module _prroi_pooling...
ninja: no work to do.
Loading extension module _prroi_pooling...
/home/zhuzhu/anaconda3/envs/pytracking/lib/python3.7/site-packages/torch/nn/functional.py:2506: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  "See the documentation of nn.Upsample for details.".format(mode))
FPS: 15.552867079669237
Tracker: atom default 0 ,  Sequence: uav_bike2
FPS: 12.340933278850803
Tracker: atom default 0 ,  Sequence: uav_bike3
FPS: 13.997377981498632
Tracker: atom default 0 ,  Sequence: uav_bird1_1
FPS: 25.896361588703243
Tracker: atom default 0 ,  Sequence: uav_bird1_2
FPS: 19.321632726340535
Tracker: atom default 0 ,  Sequence: uav_bird1_3
FPS: 18.20224524394253
Tracker: atom default 0 ,  Sequence: uav_boat1
FPS: 17.813091853676458
Tracker: atom default 0 ,  Sequence: uav_boat2
FPS: 15.937181333843958
Tracker: atom default 0 ,  Sequence: uav_boat3
FPS: 14.581508779751058
Tracker: atom default 0 ,  Sequence: uav_boat4
FPS: 13.988968325863956
Tracker: atom default 0 ,  Sequence: uav_boat5
FPS: 14.31552820225064
Tracker: atom default 0 ,  Sequence: uav_boat6
FPS: 14.094014766709591
Tracker: atom default 0 ,  Sequence: uav_boat7
FPS: 15.630099889173009
Tracker: atom default 0 ,  Sequence: uav_boat8
FPS: 16.58144160587912
Tracker: atom default 0 ,  Sequence: uav_boat9
FPS: 17.896729855946123
Tracker: atom default 0 ,  Sequence: uav_building1
FPS: 17.26361969975148
Tracker: atom default 0 ,  Sequence: uav_building2
FPS: 16.038464638950124
Tracker: atom default 0 ,  Sequence: uav_building3
FPS: 17.001612095120425
Tracker: atom default 0 ,  Sequence: uav_building4
FPS: 16.88860724473311
Tracker: atom default 0 ,  Sequence: uav_building5
FPS: 13.208389547841756
Tracker: atom default 0 ,  Sequence: uav_car1_1
FPS: 12.884910786583607
Tracker: atom default 0 ,  Sequence: uav_car1_2
FPS: 13.637050178798654
Tracker: atom default 0 ,  Sequence: uav_car1_3
FPS: 16.47417236743373
Tracker: atom default 0 ,  Sequence: uav_car10
FPS: 12.918533637377847
Tracker: atom default 0 ,  Sequence: uav_car11
FPS: 15.138679862080274
Tracker: atom default 0 ,  Sequence: uav_car12
FPS: 18.844854899022536
Tracker: atom default 0 ,  Sequence: uav_car13
FPS: 13.611199023742447
Tracker: atom default 0 ,  Sequence: uav_car14
FPS: 16.795497601706003
Tracker: atom default 0 ,  Sequence: uav_car15
FPS: 17.60971154635478
Tracker: atom default 0 ,  Sequence: uav_car16_1
FPS: 18.006560413201736
Tracker: atom default 0 ,  Sequence: uav_car16_2
FPS: 21.52928731833449
Tracker: atom default 0 ,  Sequence: uav_car17
FPS: 14.883771077844997
Tracker: atom default 0 ,  Sequence: uav_car18
FPS: 12.891138720966383
Tracker: atom default 0 ,  Sequence: uav_car1_s
FPS: 13.403854592609585
Tracker: atom default 0 ,  Sequence: uav_car2
FPS: 14.51209340684197
Tracker: atom default 0 ,  Sequence: uav_car2_s
FPS: 18.093553451117952
Tracker: atom default 0 ,  Sequence: uav_car3
FPS: 20.920079388948068
Tracker: atom default 0 ,  Sequence: uav_car3_s
FPS: 18.866010009270322
Tracker: atom default 0 ,  Sequence: uav_car4
FPS: 13.122542459551745
Tracker: atom default 0 ,  Sequence: uav_car4_s
FPS: 14.87292788599812
Tracker: atom default 0 ,  Sequence: uav_car5
FPS: 13.123920888131549
Tracker: atom default 0 ,  Sequence: uav_car6_1
FPS: 15.905125245618768
Tracker: atom default 0 ,  Sequence: uav_car6_2
FPS: 18.295301145196724
Tracker: atom default 0 ,  Sequence: uav_car6_3
FPS: 13.374666027545086
Tracker: atom default 0 ,  Sequence: uav_car6_4
FPS: 12.929125914663699
Tracker: atom default 0 ,  Sequence: uav_car6_5
FPS: 14.016775058057677
Tracker: atom default 0 ,  Sequence: uav_car7
FPS: 15.731279272358755
Tracker: atom default 0 ,  Sequence: uav_car8_1
FPS: 13.13203749613079
Tracker: atom default 0 ,  Sequence: uav_car8_2
FPS: 13.170991011333458
Tracker: atom default 0 ,  Sequence: uav_car9
FPS: 16.10489335661624
Tracker: atom default 0 ,  Sequence: uav_group1_1
FPS: 12.350973458574384
Tracker: atom default 0 ,  Sequence: uav_group1_2
FPS: 15.75623018460669
Tracker: atom default 0 ,  Sequence: uav_group1_3
FPS: 12.809226059646408
Tracker: atom default 0 ,  Sequence: uav_group1_4
FPS: 13.022193012506783
Tracker: atom default 0 ,  Sequence: uav_group2_1
FPS: 24.265265154321686
Tracker: atom default 0 ,  Sequence: uav_group2_2
FPS: 22.692301062386136
Tracker: atom default 0 ,  Sequence: uav_group2_3
FPS: 18.915077628808476
Tracker: atom default 0 ,  Sequence: uav_group3_1
FPS: 12.908029749659212
Tracker: atom default 0 ,  Sequence: uav_group3_2
FPS: 15.46210892394108
Tracker: atom default 0 ,  Sequence: uav_group3_3
FPS: 17.910556850154425
Tracker: atom default 0 ,  Sequence: uav_group3_4
FPS: 20.662544363257073
Tracker: atom default 0 ,  Sequence: uav_person1
FPS: 11.52524713817831
Tracker: atom default 0 ,  Sequence: uav_person10
FPS: 13.50419502076276
Tracker: atom default 0 ,  Sequence: uav_person11
FPS: 13.28694456752528
Tracker: atom default 0 ,  Sequence: uav_person12_1
FPS: 13.989250619974044
Tracker: atom default 0 ,  Sequence: uav_person12_2
FPS: 17.347241039541505
Tracker: atom default 0 ,  Sequence: uav_person13
FPS: 12.97812402270601
Tracker: atom default 0 ,  Sequence: uav_person14_1
FPS: 18.52467264757104
Tracker: atom default 0 ,  Sequence: uav_person14_2
FPS: 13.88615129461356
Tracker: atom default 0 ,  Sequence: uav_person14_3
FPS: 16.75607336459585
Tracker: atom default 0 ,  Sequence: uav_person15
FPS: 20.94138848552354
Tracker: atom default 0 ,  Sequence: uav_person16
FPS: 13.767360147804018
Tracker: atom default 0 ,  Sequence: uav_person17_1
FPS: 13.282334375876573
Tracker: atom default 0 ,  Sequence: uav_person17_2
FPS: 16.019002818626255
Tracker: atom default 0 ,  Sequence: uav_person18
FPS: 15.048773818266397
Tracker: atom default 0 ,  Sequence: uav_person19_1
FPS: 13.17366164629562
Tracker: atom default 0 ,  Sequence: uav_person19_2
FPS: 13.179767104227835
Tracker: atom default 0 ,  Sequence: uav_person19_3
FPS: 14.05881708799293
Tracker: atom default 0 ,  Sequence: uav_person1_s
FPS: 16.648982250532406
Tracker: atom default 0 ,  Sequence: uav_person2_1
FPS: 12.305198196628972
Tracker: atom default 0 ,  Sequence: uav_person2_2
FPS: 14.666007954893026
Tracker: atom default 0 ,  Sequence: uav_person20
FPS: 14.960058130237913
Tracker: atom default 0 ,  Sequence: uav_person21
FPS: 12.080122969423373
Tracker: atom default 0 ,  Sequence: uav_person22
FPS: 12.653201040451105
Tracker: atom default 0 ,  Sequence: uav_person23
FPS: 12.037050671733653
Tracker: atom default 0 ,  Sequence: uav_person2_s
FPS: 11.907740043520004
Tracker: atom default 0 ,  Sequence: uav_person3
FPS: 12.555642277519498
Tracker: atom default 0 ,  Sequence: uav_person3_s
FPS: 13.374827230578129
Tracker: atom default 0 ,  Sequence: uav_person4_1
FPS: 17.994959845129333
Tracker: atom default 0 ,  Sequence: uav_person4_2
FPS: 11.25448020292777
Tracker: atom default 0 ,  Sequence: uav_person5_1
FPS: 11.552936418646082
Tracker: atom default 0 ,  Sequence: uav_person5_2
FPS: 11.538577668278895
Tracker: atom default 0 ,  Sequence: uav_person6
FPS: 12.096996875733801
Tracker: atom default 0 ,  Sequence: uav_person7_1
FPS: 18.809579397705548
Tracker: atom default 0 ,  Sequence: uav_person7_2
FPS: 11.825422613289515
Tracker: atom default 0 ,  Sequence: uav_person8_1
FPS: 12.288598511010692
Tracker: atom default 0 ,  Sequence: uav_person8_2
FPS: 10.40009629772226
Tracker: atom default 0 ,  Sequence: uav_person9
FPS: 6.280039353994421
Tracker: atom default 0 ,  Sequence: uav_truck1
FPS: 15.308288451019648
Tracker: atom default 0 ,  Sequence: uav_truck2
FPS: 7.724032633465111
Tracker: atom default 0 ,  Sequence: uav_truck3
FPS: 8.17653401665531
Tracker: atom default 0 ,  Sequence: uav_truck4_1
FPS: 10.457750492125374
Tracker: atom default 0 ,  Sequence: uav_truck4_2
FPS: 10.432150321703007
Tracker: atom default 0 ,  Sequence: uav_uav1_1
FPS: 19.298286344937917
Tracker: atom default 0 ,  Sequence: uav_uav1_2
FPS: 19.91448410695476
Tracker: atom default 0 ,  Sequence: uav_uav1_3
FPS: 18.56974354757066
Tracker: atom default 0 ,  Sequence: uav_uav2
FPS: 14.407311229185144
Tracker: atom default 0 ,  Sequence: uav_uav3
FPS: 10.73631313710876
Tracker: atom default 0 ,  Sequence: uav_uav4
FPS: 14.606850187147947
Tracker: atom default 0 ,  Sequence: uav_uav5
FPS: 11.44107359473324
Tracker: atom default 0 ,  Sequence: uav_uav6
FPS: 10.873437707401798
Tracker: atom default 0 ,  Sequence: uav_uav7
FPS: 14.459721205350634
Tracker: atom default 0 ,  Sequence: uav_uav8
FPS: 18.799643694638043
Tracker: atom default 0 ,  Sequence: uav_wakeboard1
FPS: 11.280216396954726
Tracker: atom default 0 ,  Sequence: uav_wakeboard10
FPS: 12.002140866259124
Tracker: atom default 0 ,  Sequence: uav_wakeboard2
FPS: 15.691527156715846
Tracker: atom default 0 ,  Sequence: uav_wakeboard3
FPS: 19.57336049359283
Tracker: atom default 0 ,  Sequence: uav_wakeboard4
FPS: 12.543842609466745
Tracker: atom default 0 ,  Sequence: uav_wakeboard5
FPS: 9.389698892653701
Tracker: atom default 0 ,  Sequence: uav_wakeboard6
FPS: 10.986393742594911
Tracker: atom default 0 ,  Sequence: uav_wakeboard7
FPS: 9.914664371278853
Tracker: atom default 0 ,  Sequence: uav_wakeboard8
FPS: 15.631635419583796
Tracker: atom default 0 ,  Sequence: uav_wakeboard9
FPS: 12.70858852063887
Done
程序共运行：17877.000000秒

在这里插入图片描述在这里插入图片描述在这里插入图片描述

UAV:ATOM
AUC:61.90
OP50:74.61
OP75:53.96
Precision:82.79
Norm Precision:75.93

run_experiment.py debug=0 atom default atom_otb range(3)(三个atom跟踪器)
结果保存在pytracking.tracking_results_me_tracker_num3.atom.default_000/001/002
UAV | AUC | OP50 | OP75 | Precision | Norm Precision |
ATOM | 61.65 | 74.54 | 53.91 | 82.59 | 76.05 |

朱朱呀
关注
  • 1
    点赞
  • 踩
  • 2
    收藏
    觉得还不错? 一键收藏
  • 2
    评论
专栏目录
keras resnet 训练自己的数据
04-30
keras resnet 迁移训练数据 和 读取数据
改写ResNet-18训练自己的数据集
小膜仙的博客
03-01 3375
数据集准备 之前做好的 训练过程结果 十类，45.2174的准确率
深度学习基于Resnet18的图像多分类--训练自己的数据集(超详细 含源码)
热门推荐
weixin_49345320的博客
06-22 1万+
深度学习基于Resnet18的图像多分类--训练自己的数据集(超详细 含源码)
pytorch基于resnet18预训练模型用于自己的训练数据集进行迁移学习
booklijian的专栏
07-08 1万+
本文记录利用resnet18预训练模型进行迁移学习，在自己的训练数据集上进行重新训练。相关代码重点部分分别介绍如下： model=torchvision.models.resnet18(pretrained=True) num_features=model.fc.in_features model.fc=nn.Linear(num_features,num_classes) model=model.to(device) 说明：上述代码首先加载resnet18预训练模型，然后根据训练数据集中的分类数量
UAV-UAV(U2U)、UAV-ground user(U2G)通信信道模型
thompson的博客
02-18 1276
摘自 Zhao H, Wang H, Wu W, et al. Deployment algorithms for UAV airborne networks toward on-demand coverage[J]. IEEE Journal on Selected Areas in Communications, 2018, 36(9): 2015-2031.
atom_default resnet18_vggmconv1 网络模型
03-23
"atom_default.pth"和"resnet18_vggmconv1.pth"是两个权重文件，可能是训练好的网络模型的权重。".pth"文件是PyTorch库中常用的模型权重保存格式。"atom_default.pth"可能代表一个名为"atom_default"的模型的默认...
dlib_face_recognition_resnet_model_v1.dat.zip_ResNet_dlib_face r
07-15
dlib_face_recognition_resnet_model_v1.dat是dlib库提供的预训练模型文件，它包含了经过大量面部图像数据训练后的权重和参数。使用这个模型，我们可以将输入的人脸图像转化为128维的特征向量，这些特征向量能够有效...
resnet_v2(预训练模型).tar.gz
09-24
当神经网络具有许多参数时，它们的...但是，这意味着必须对非常大的数据集进行训练。由于从头开始训练模型可能是一个非常耗费计算量的过程，需要几天甚至几周的时间，因此，这里提供了各种预先训练的模型供大家下载！
ResNet18_resnet18_resnet18网络_
10-02
用户可以利用这个文件直接实例化ResNet18模型，并进行训练、验证或测试。 `ResNet18结构1.png`和`ResNet结构2.png`文件可能是ResNet18网络结构的可视化图，帮助理解模型的布局。通常，这些图会展示每个残差块的连接...
inception_resnet_v2_2016_08_30预训练模型
08-11
4. **训练**: 使用新数据集进行训练，可能需要调整学习率、优化器和批大小等超参数。 5. **评估和优化**: 训练后，评估模型在验证集上的性能，并根据结果进行超参数调整或网络结构调整。 Inception_ResNet_v2的预...
pytorch+resnet18实现长尾数据集分类(一)
12-20
实验基于论文： Class-Balanced Loss Based on Effective Number of Samples Class-balanced-loss代码地址：https://github.com/vandit15/Class-balanced-loss-pytorch resnet18代码参考链接：https://blog.csdn.net/sunqiande88/article/details/80100891 制作数据集 论文中通过公式n=niuin = n_iu^in=ni​ui，iii为类索引．制作长尾cifar10数据集．以下代码以不均匀比例100为例．也可以通过
pytorch-resnet18和resnet50官方预训练模型
08-22
pytroch官网提供的预训练模型：resnet18：resnet18-5c106cde.pth和resnet50：resnet50-19c8e357.pth（两个文件打包在一起）
用PyTorch在一个物体数据库上训练ResNet
08-22
最近几年，在处理语音、图像识别和语言处理等问题上，深度学习有着十分不错的表现。在所有类型神经网络中，卷积神经网络中研究最为深入的。在过去因为没有大数据与高性能的计算设备支撑，想要在不过拟合进行高性能卷积神经网络训练是难以想象的。如今卷积神经网络的飞速发展，离不开出现了ImageNet这样的数据，还有大幅提升的GPU计算性能。尽管各种 CNN 模型仍继续在多种计算机视觉应用中进一步推进当前最佳的表现，但在理解这些系统的工作方式和如此有效的原因上的进展仍还有限。这个问题已经引起了很多研究者的兴趣，为此也涌现出了很多用于理解 CNN 的方法。本课题主要研究了卷积神经网络ResNet，在pytorch平台进行对ResNet34神经网络的训练，得到达到识别精度的神经网络，并对设计的神经网络模型进行评估和测试。首先从已有的图库中crop处理做一个5*100的图像数据库，并搭建卷积神经网络研究平台PyTorch的Window平台，按要求训练出具有识别精度的神经网络，通过设计算法来测试模型的识别率，并将把模型的卷积核可视化。
pytracking/pytracking/evaluation/tracker.py
10-31
此为用pysot里面的vot功能DIMP50要用到的修改后的tracker.py文件
使用PyTorch搭建ResNet18网络并使用CIFAR10数据集训练测试
TTTSEP9TH2244的博客
02-25 8810
ResNet18网络结构 具体ResNet原理细节这里不多做描述，直接上代码 这里并未采用BasicBlock和BottleNeck复现ResNet18 model.py网络模型部分： import torch import torch.nn as nn from torch.nn import functional as F """ 把ResNet18的残差卷积单元作为一个Block，这里分为两种：一种是CommonBlock，另一种是SpecialBlock，最后由ResNet18统筹调度 其中S
Pytorch入门（五）使用ResNet-18网络训练常规状态下的CIFAR10数据集
一个非常Cool的人
06-06 5367
本文介绍了如何使用Pytorch+ResNet-18+CIFAR-10进行深度学习训练，并使用训练好的权重进行预测，ResNet是一个非常优秀的神经网络，ResNet 在图像分类任务中表现出色。
UAV通信中的天线建模
thompson的博客
05-04 506
文章目录12 1 来自 : Mozaffari M , Saad W , Bennis M , et al. Ef?cient Deployment of Multiple Unmanned Aerial Vehicles for Optimal Wireless Coverage[J]. IEEE Communications Letters, 2016:1-1. 2 B. Galkin, ...
UAV和基站通信的天线模型
thompson的博客
07-12 385
UAV和基站通信的天线模型
基于pytorch实现Resnet对本地数据集的训练
openabox的博客
05-15 3741
本文是使用pycharm下的pytorch框架编写一个训练本地数据集的Resnet深度学习模型，其一共有两百行代码左右，分成mian.py、network.py、dataset.py以及train.py文件，功能是对本地的数据集进行分类。本文介绍逻辑是总分形式，即首先对总流程进行一个概括，然后分别介绍每个流程中的实现过程（代码+流程图+文字的介绍）。
resnet18训练自己的数据集
最新发布
09-14
您可以使用PyTorch来训练自己的数据集，使用ResNet-18模型。以下是一个基本的训练脚本示例： ```python import torch import torch.nn as nn import torch.optim as optim from torchvision.models import resnet18...
评论 2
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值