这是一个yolo3-tf2的源码,可以用于训练自己的模型。

Overview

YOLOV3:You Only Look Once目标检测模型在Tensorflow2当中的实现


目录

  1. 性能情况 Performance
  2. 所需环境 Environment
  3. 文件下载 Download
  4. 训练步骤 How2train
  5. 预测步骤 How2predict
  6. 评估步骤 How2eval
  7. 参考资料 Reference

性能情况

训练数据集 权值文件名称 测试数据集 输入图片大小 mAP 0.5:0.95 mAP 0.5
COCO-Train2017 yolo_weights.h5 COCO-Val2017 416x416 38.1 66.8

所需环境

tensorflow==2.2.0

文件下载

训练所需的yolo_weights.pth可以在百度云下载。
链接: https://pan.baidu.com/s/1URJJiPUYjiWzitU0nytBnA 提取码: qvxs

VOC数据集下载地址如下:
VOC2007+2012训练集
链接: https://pan.baidu.com/s/16pemiBGd-P9q2j7dZKGDFA 提取码: eiw9

VOC2007测试集
链接: https://pan.baidu.com/s/1BnMiFwlNwIWG9gsd4jHLig 提取码: dsda

训练步骤

a、数据集的准备

本文使用VOC格式进行训练,训练前需要自己制作好数据集,如果没有自己的数据集,可以通过Github连接下载VOC12+07的数据集尝试下。
训练前将标签文件放在VOCdevkit文件夹下的VOC2007文件夹下的Annotation中。
训练前将图片文件放在VOCdevkit文件夹下的VOC2007文件夹下的JPEGImages中。

b、数据集的处理

在完成数据集的摆放之后,我们需要对数据集进行下一步的处理,目的是获得训练用的2007_train.txt以及2007_val.txt,需要用到根目录下的voc_annotation.py。
voc_annotation.py里面有一些参数需要设置。第一次训练可以仅修改classes_path,classes_path用于指向检测类别所对应的txt。
训练自己的数据集时,可以自己建立一个cls_classes.txt,里面写自己所需要区分的类别。
model_data/cls_classes.txt文件内容为:

cat
dog
...

c、开始网络训练

通过voc_annotation.py我们已经生成了2007_train.txt以及2007_val.txt,此时我们可以开始训练了。训练的参数较多,大家可以在下载库后仔细看注释,其中最重要的部分依然是train.py里的classes_path。
classes_path用于指向检测类别所对应的txt,这个txt和voc_annotation.py里面的txt一样!训练自己的数据集必须要修改!
修改完classes_path后就可以运行train.py开始训练了,在训练多个epoch后,权值会生成在logs文件夹中。

d、训练结果预测

训练结果预测需要用到两个文件,分别是yolo.py和predict.py。我们首先需要去yolo.py里面修改model_path以及classes_path,这两个参数必须要修改。
model_path指向训练好的权值文件,在logs文件夹里。
classes_path指向检测类别所对应的txt。

完成修改后就可以运行predict.py进行检测了。运行后输入图片路径即可检测。

预测步骤

a、使用预训练权重

  1. 下载完库后解压,在百度网盘下载yolo_weights.pth,放入model_data,运行predict.py,输入
img/street.jpg
  1. 在predict.py里面进行设置可以进行fps测试和video视频检测。

b、使用自己训练的权重

  1. 按照训练步骤训练。
  2. 在yolo.py文件里面,在如下部分修改model_path和classes_path使其对应训练好的文件;model_path对应logs文件夹下面的权值文件,classes_path是model_path对应分的类
_defaults = {
    #--------------------------------------------------------------------------#
    #   使用自己训练好的模型进行预测一定要修改model_path和classes_path!
    #   model_path指向logs文件夹下的权值文件,classes_path指向model_data下的txt
    #   如果出现shape不匹配,同时要注意训练时的model_path和classes_path参数的修改
    #--------------------------------------------------------------------------#
    "model_path"        : 'model_data/yolo_weight.h5',
    "classes_path"      : 'model_data/coco_classes.txt',
    #---------------------------------------------------------------------#
    #   anchors_path代表先验框对应的txt文件,一般不修改。
    #   anchors_mask用于帮助代码找到对应的先验框,一般不修改。
    #---------------------------------------------------------------------#
    "anchors_path"      : 'model_data/yolo_anchors.txt',
    "anchors_mask"      : [[6, 7, 8], [3, 4, 5], [0, 1, 2]],
    #---------------------------------------------------------------------#
    #   输入图片的大小,必须为32的倍数。
    #---------------------------------------------------------------------#
    "input_shape"       : [416, 416],
    #---------------------------------------------------------------------#
    #   只有得分大于置信度的预测框会被保留下来
    #---------------------------------------------------------------------#
    "confidence"        : 0.5,
    #---------------------------------------------------------------------#
    #   非极大抑制所用到的nms_iou大小
    #---------------------------------------------------------------------#
    "nms_iou"           : 0.3,
    "max_boxes"         : 100,
    #---------------------------------------------------------------------#
    #   该变量用于控制是否使用letterbox_image对输入图像进行不失真的resize,
    #   在多次测试后,发现关闭letterbox_image直接resize的效果更好
    #---------------------------------------------------------------------#
    "letterbox_image"   : False,
}
  1. 运行predict.py,输入
img/street.jpg
  1. 在predict.py里面进行设置可以进行fps测试和video视频检测。

评估步骤

  1. 本文使用VOC格式进行评估。
  2. 如果在训练前已经运行过voc_annotation.py文件,代码会自动将数据集划分成训练集、验证集和测试集。如果想要修改测试集的比例,可以修改voc_annotation.py文件下的trainval_percent。trainval_percent用于指定(训练集+验证集)与测试集的比例,默认情况下 (训练集+验证集):测试集 = 9:1。train_percent用于指定(训练集+验证集)中训练集与验证集的比例,默认情况下 训练集:验证集 = 9:1。
  3. 利用voc_annotation.py划分测试集后,前往get_map.py文件修改classes_path,classes_path用于指向检测类别所对应的txt,这个txt和训练时的txt一样。评估自己的数据集必须要修改。
  4. 运行get_map.py即可获得评估结果,评估结果会保存在map_out文件夹中。

Reference

https://github.com/qqwweee/keras-yolo3
https://github.com/eriklindernoren/PyTorch-YOLOv3
https://github.com/BobLiu20/YOLOv3_PyTorch

You might also like...
Comments
  • W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated. 	 [[{{node PyFunc}}]]

    W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated. [[{{node PyFunc}}]]

    f.write("%s %s %s %s %s %s\n" % (predicted_class, score[:6], str(int(left)), str(int(top)), str(int(right)),str(int(bottom))))
    

    OverflowError: cannot convert float infinity to integer 2022-10-17 08:14:29.544536: W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Failed precondition: Python interpreter state is not initialized. The process may be terminated. [[{{node PyFunc}}]] upup 这个报错是什么意思呢

    opened by 11wswhl 2
  • Resource exhausted: OOM when allocating tensor with shape[255,256,256,69] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc

    Resource exhausted: OOM when allocating tensor with shape[255,256,256,69] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc

    I meet an error:

    2022-01-30 10:37:36.925076: I tensorflow/core/common_runtime/bfc_allocator.cc:923] total_region_allocated_bytes_: 1074790400 memory_limit_: 3059430195 available bytes: 1984639795 curr_region_allocation_bytes_: 1073741824 2022-01-30 10:37:36.925383: I tensorflow/core/common_runtime/bfc_allocator.cc:929] Stats: Limit: 3059430195 InUse: 621716992 MaxInUse: 621716992 NumAllocs: 1595 MaxAllocSize: 33554432

    2022-01-30 10:37:36.925822: W tensorflow/core/common_runtime/bfc_allocator.cc:424] **********************************************************__________________________________________ 2022-01-30 10:37:36.926021: W tensorflow/core/framework/op_kernel.cc:1651] OP_REQUIRES failed at cwise_ops_common.cc:82 : Resource exhausted: OOM when allocating tensor with shape[255,256,256,69] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc

    Could you please help me solve this problem

    opened by Linamiao-1998 1
  • 您好,我可能发现了一个在yolo_traing.py中的bug

    您好,我可能发现了一个在yolo_traing.py中的bug

    是这样的,因为我的数据长宽大概为256,640这样的,所以我准备设置将训练的inputshape也设置为256,640,但训练时损失会很快变为负数。 而当我全程使用正方形训练和测试都没有问题.使用inputshape为正方形训练,然后设置inputshape为矩形测试也没有问题 经过一路排查,应该是将y_true的x,y转为x,y的偏移量时grid_shapes[l]的长,宽顺序反了: image

    将其中的grid_shapes[l][:]改为grid_shapes[l][::-1]我的训练就正常了。 因为正方形长宽高相等,所以不影响,而矩形这样会导致偏移量有大于1的值产生,在使用keras.binary_crossentropy计算损失时,label大于1则会产生负损失。 keras.binary_crossentropy(label , logits, from_logits=True)的计算公式如下: x = logits, z = labels loss = max(x, 0) - x * z + log(1 + exp(-abs(x)))

    opened by kill2013110 2
  • 请问有尝试过将该模型保存为savedmodel格式的吗?我保存成这样后再读取就会出错

    请问有尝试过将该模型保存为savedmodel格式的吗?我保存成这样后再读取就会出错

    使用savedmodel是因为要后续准备冻结模型。 函数用的tf.keras.model.save_model()和tf.keras.model.load_model() 分析了一下,是因为模型中包含了自定义的Lambda层,所以报错了, 对于lambda层我准备转为tf.function, 应该可以解决问题, 您有什么好的建议吗

    opened by kill2013110 4
Releases(v3.2)
  • v3.2(Jul 16, 2022)

    重要更新

    • 增加了训练时评估,可以在train.py进行开关或者调整评估周期。
    • 更新了评估代码,可以设置计算Recall和Precision的门限。
    • 在summary.py新增网络各类参数计算。
    • 增加保存权值的方法,loss最低、最近一次等。
    • 新增大量注释。
    Source code(tar.gz)
    Source code(zip)
  • v3.1(Feb 23, 2022)

    重要更新

    • 修改了loss组成,使得分类、目标、回归loss的比例合适。
    • 支持step、cos学习率下降法。
    • 支持adam、sgd优化器选择。
    • 支持学习率根据batch_size自适应调整。
    • 支持不同预测模式的选择,单张图片预测、文件夹预测、视频预测、图片裁剪、heatmap、各个种类目标数量计算。
    • 更新summary.py文件,用于观看网络结构。
    • 增加了多GPU训练。
    Source code(tar.gz)
    Source code(zip)
  • v2.0(Feb 21, 2022)

    重要更新

    • 更新train.py文件,增加了大量的注释,增加多个可调整参数。
    • 更新predict.py文件,增加了大量的注释,增加fps、视频预测、批量预测等功能。
    • 更新yolo.py文件,增加了大量的注释,增加先验框选择、置信度、非极大抑制等参数。
    • 合并get_dr_txt.py、get_gt_txt.py和get_map.py文件,通过一个文件来实现数据集的评估。
    • 更新voc_annotation.py文件,增加多个可调整参数。
    • 更新kmeans_for_anchors.py文件,用于计算先验框的大小。
    • 更新summary.py文件,用于观看网络结构。
    Source code(tar.gz)
    Source code(zip)
Owner
Bubbliiiing
Bubbliiiing
Official code for the paper: Deep Graph Matching under Quadratic Constraint (CVPR 2021)

QC-DGM This is the official PyTorch implementation and models for our CVPR 2021 paper: Deep Graph Matching under Quadratic Constraint. It also contain

Quankai Gao 55 Nov 14, 2022
fastgradio is a python library to quickly build and share gradio interfaces of your trained fastai models.

fastgradio is a python library to quickly build and share gradio interfaces of your trained fastai models.

Ali Abdalla 34 Jan 05, 2023
Reinfore learning tool box, contains trpo, a3c algorithm for continous action space

RL_toolbox all the algorithm is running on pycharm IDE, or the package loss error may exist. implemented algorithm: trpo a3c a3c:for continous action

yupei.wu 44 Oct 10, 2022
TransFGU: A Top-down Approach to Fine-Grained Unsupervised Semantic Segmentation

TransFGU: A Top-down Approach to Fine-Grained Unsupervised Semantic Segmentation Zhaoyun Yin, Pichao Wang, Fan Wang, Xianzhe Xu, Hanling Zhang, Hao Li

DamoCV 25 Dec 16, 2022
Code for the ECCV2020 paper "A Differentiable Recurrent Surface for Asynchronous Event-Based Data"

A Differentiable Recurrent Surface for Asynchronous Event-Based Data Code for the ECCV2020 paper "A Differentiable Recurrent Surface for Asynchronous

Marco Cannici 21 Oct 05, 2022
All of the figures and notebooks for my deep learning book, for free!

"Deep Learning - A Visual Approach" by Andrew Glassner This is the official repo for my book from No Starch Press. Ordering the book My book is called

Andrew Glassner 227 Jan 04, 2023
Inferring Lexicographically-Ordered Rewards from Preferences

Inferring Lexicographically-Ordered Rewards from Preferences Code author: Alihan Hüyük ([e

Alihan Hüyük 1 Feb 13, 2022
A coin flip game in which you can put the amount of money below or equal to 1000 and then choose heads or tail

COIN_FLIPPY ##This is a simple example package. You can use Github-flavored Markdown to write your content. Coinflippy A coin flip game in which you c

2 Dec 26, 2021
A Multi-modal Perception Tracker (MPT) for speaker tracking using both audio and visual modalities

MPT A Multi-modal Perception Tracker (MPT) for speaker tracking using both audio and visual modalities. Implementation for our AAAI 2022 paper: Multi-

yidiLi 4 May 08, 2022
PyTorch implementation for paper "Full-Body Visual Self-Modeling of Robot Morphologies".

Full-Body Visual Self-Modeling of Robot Morphologies Boyuan Chen, Robert Kwiatkowskig, Carl Vondrick, Hod Lipson Columbia University Project Website |

Boyuan Chen 32 Jan 02, 2023
A modular, research-friendly framework for high-performance and inference of sequence models at many scales

T5X T5X is a modular, composable, research-friendly framework for high-performance, configurable, self-service training, evaluation, and inference of

Google Research 1.1k Jan 08, 2023
SmartSim Infrastructure Library.

Home Install Documentation Slack Invite Cray Labs SmartSim SmartSim makes it easier to use common Machine Learning (ML) libraries like PyTorch and Ten

Cray Labs 139 Jan 01, 2023
🥇 LG-AI-Challenge 2022 1위 솔루션 입니다.

LG-AI-Challenge-for-Plant-Classification Dacon에서 진행된 농업 환경 변화에 따른 작물 병해 진단 AI 경진대회 에 대한 코드입니다. (colab directory에 코드가 잘 정리 되어있습니다.) Requirements python

siwooyong 10 Jun 30, 2022
Discriminative Condition-Aware PLDA

DCA-PLDA This repository implements the Discriminative Condition-Aware Backend described in the paper: L. Ferrer, M. McLaren, and N. Brümmer, "A Speak

Luciana Ferrer 31 Aug 05, 2022
[2021 MultiMedia] CONQUER: Contextual Query-aware Ranking for Video Corpus Moment Retrieval

CONQUER: Contexutal Query-aware Ranking for Video Corpus Moment Retreival PyTorch implementation of CONQUER: Contexutal Query-aware Ranking for Video

Hou zhijian 23 Dec 26, 2022
Get 2D point positions (e.g., facial landmarks) projected on 3D mesh

points2d_projection_mesh Input 2D points (e.g. facial landmarks) on an image Camera parameters (extrinsic and intrinsic) of the image Aligned 3D mesh

5 Dec 08, 2022
SEOVER: Sentence-level Emotion Orientation Vector based Conversation Emotion Recognition Model

SEOVER-Master This code is the implementation of paper: SEOVER: Sentence-level Emotion Orientation Vector based Conversation Emotion Recognition Model

4 Feb 24, 2022
DuBE: Duple-balanced Ensemble Learning from Skewed Data

DuBE: Duple-balanced Ensemble Learning from Skewed Data "Towards Inter-class and Intra-class Imbalance in Class-imbalanced Learning" (IEEE ICDE 2022 S

6 Nov 12, 2022
A Dynamic Residual Self-Attention Network for Lightweight Single Image Super-Resolution

DRSAN A Dynamic Residual Self-Attention Network for Lightweight Single Image Super-Resolution Karam Park, Jae Woong Soh, and Nam Ik Cho Environments U

4 May 10, 2022
Unofficial TensorFlow implementation of the Keyword Spotting Transformer model

Keyword Spotting Transformer This is the unofficial TensorFlow implementation of the Keyword Spotting Transformer model. This model is used to train o

Intelligent Machines Limited 8 May 11, 2022