CenterNet:Objects as Points目标检测模型在Pytorch当中的实现

Related tags

Deep Learningcenternet-pytorch
Overview

CenterNet:Objects as Points目标检测模型在Pytorch当中的实现

目录

  1. 性能情况 Performance
  2. 所需环境 Environment
  3. 注意事项 Attention
  4. 文件下载 Download
  5. 预测步骤 How2predict
  6. 训练步骤 How2train
  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
VOC07+12 centernet_resnet50_voc.pth VOC-Test07 512x512 - 77.1
COCO-Train2017 centernet_hourglass_coco.pth COCO-Val2017 512x512 38.4 56.8

所需环境

torch==1.2.0

注意事项

代码中的centernet_resnet50_voc.pth是使用voc数据集训练的。
代码中的centernet_hourglass_coco.pth是使用coco数据集训练的。
注意不要使用中文标签，文件夹中不要有空格！
在训练前需要务必在model_data下新建一个txt文档，文档中输入需要分的类，在train.py中将classes_path指向该文件。

文件下载

训练所需的centernet_resnet50_voc.pth、centernet_hourglass_coco.pth可在百度网盘中下载。
链接: https://pan.baidu.com/s/1QBBgRb_TH8kJdSCQGgcXmQ 提取码: phnc

centernet_resnet50_voc.pth是voc数据集的权重。
centernet_hourglass_coco.pth是coco数据集的权重。

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centernet_resnet50_voc.pth或者centernet_hourglass_coco.pth，放入model_data，运行predict.py，输入
img/street.jpg
  1. 利用video.py可进行摄像头检测。

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 = {
    "model_path"        : 'model_data/centernet_resnet50_voc.pth',
    "classes_path"      : 'model_data/voc_classes.txt',
    # "model_path"        : 'model_data/centernet_hourglass_coco.h5',
    # "classes_path"      : 'model_data/coco_classes.txt',
    "backbone"          : "resnet50",
    "image_size"        : [512,512,3],
    "confidence"        : 0.3,
    # backbone为resnet50时建议设置为True
    # backbone为hourglass时建议设置为False
    # 也可以根据检测效果自行选择
    "nms"               : True,
    "nms_threhold"      : 0.3,
    "cuda"              : True
}
  1. 运行predict.py，输入
img/street.jpg
  1. 利用video.py可进行摄像头检测。

训练步骤

  1. 本文使用VOC格式进行训练。
  2. 训练前将标签文件放在VOCdevkit文件夹下的VOC2007文件夹下的Annotation中。
  3. 训练前将图片文件放在VOCdevkit文件夹下的VOC2007文件夹下的JPEGImages中。
  4. 在训练前利用voc2centernet.py文件生成对应的txt。
  5. 再运行根目录下的voc_annotation.py，运行前需要将classes改成你自己的classes。注意不要使用中文标签，文件夹中不要有空格！
classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]
  1. 此时会生成对应的2007_train.txt，每一行对应其图片位置及其真实框的位置。
  2. 在训练前需要务必在model_data下新建一个txt文档，文档中输入需要分的类，在train.py中将classes_path指向该文件，示例如下：
classes_path = 'model_data/new_classes.txt'

model_data/new_classes.txt文件内容为：

cat
dog
...
  1. 运行train.py即可开始训练。

mAP目标检测精度计算更新

更新了get_gt_txt.py、get_dr_txt.py和get_map.py文件。
get_map文件克隆自https://github.com/Cartucho/mAP
具体mAP计算过程可参考：https://www.bilibili.com/video/BV1zE411u7Vw

Reference

https://github.com/xuannianz/keras-CenterNet
https://github.com/see--/keras-centernet
https://github.com/xingyizhou/CenterNet

Comments
  • map指标

    map指标

    B导，我在使用get_map.py的时候，您的初始设置confidence为0.02，我正常得到map结果，但是我像其他网络一样把confidence修改成为0.001以后就得不到map结果了，这是为什么呢？ 还有就是想问一下，在计算voc的map时，confidence都应该设置为很低，所以是不是0.02和0.001的效果相似？谢谢b导

    opened by ChristmasLee 2
  • 训练没有归一化，预测却有归一化，是不是有问题？

    训练没有归一化，预测却有归一化，是不是有问题？

    训练时候加载数据是dataloader.py 222行，是没有对图片做mean和std归一化的，但预测时predict.py -> centernet.py -> util/util.py -> preprocess_input里却对图片做了mean、std归一化，这应该有问题吧？

    opened by seven-linglx 2
  • 显示no mudule named 'past'

    显示no mudule named 'past'

    Traceback (most recent call last): File "train.py", line 15, in from utils.callbacks import LossHistory File "/root/centernet/centernet-pytorch-main/utils/callbacks.py", line 9, in from torch.utils.tensorboard import SummaryWriter File "/root/.local/lib/python3.7/site-packages/torch/utils/tensorboard/init.py", line 6, in from .writer import FileWriter, SummaryWriter # noqa F401 File "/root/.local/lib/python3.7/site-packages/torch/utils/tensorboard/writer.py", line 18, in from ._convert_np import make_np File "/root/.local/lib/python3.7/site-packages/torch/utils/tensorboard/_convert_np.py", line 12, in from caffe2.python import workspace File "/root/.local/lib/python3.7/site-packages/caffe2/python/workspace.py", line 15, in from past.builtins import basestring

    opened by buloseshi 1
  • 请问我改mobilenetv3的时候运行到第7批次就自动停止了是怎么回事呢

    请问我改mobilenetv3的时候运行到第7批次就自动停止了是怎么回事呢

    Finish Validation 0%| | 0/119 [00:00<?, ?it/s]Get map. 0%| | 0/119 [00:00<?, ?it/s] Traceback (most recent call last): File "/home/linux/data2/sun/centernet-pytorch-main/train.py", line 491, in epoch_step, epoch_step_val, gen, gen_val, UnFreeze_Epoch, Cuda, fp16, scaler, backbone, save_period, save_dir, local_rank) File "/home/linux/data2/sun/centernet-pytorch-main/utils/utils_fit.py", line 161, in fit_one_epoch eval_callback.on_epoch_end(epoch + 1, model_train) File "/home/linux/data2/sun/centernet-pytorch-main/utils/callbacks.py", line 211, in on_epoch_end self.get_map_txt(image_id, image, self.class_names, self.map_out_path) File "/home/linux/data2/sun/centernet-pytorch-main/utils/callbacks.py", line 145, in get_map_txt outputs = decode_bbox(outputs[0], outputs[1], outputs[2], self.confidence, self.cuda) IndexError: list index out of range

    opened by sunsn1997 2
  • 第一次尝试的新手提问

    第一次尝试的新手提问

    image

    按照readme文档中的步骤 1 已解压VOC数据集至项目根目录，pth文件至model_data目录 2 已修改voc_annotation.py 中的annotation_mode为2 3 运行train.py

    环境 pytorch1.2 + cuda10.0 +python3.6 ，Ubuntu 刚开始是使用的高版本torch和python，然后也尝试了python3.6+ torch1.2的环境,出现一样的问题

    opened by Xie-Muxi 1
Releases(v3.0)

  • v3.0(Apr 22, 2022)

    重要更新

    • 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(Mar 4, 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centernet.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summary.py文件，用于观看网络结构。
    Source code(tar.gz)
    Source code(zip)

  • v1.0(Dec 17, 2020)

Owner
Bubbliiiing
Bubbliiiing
GitHub Repository
A Transformer-Based Siamese Network for Change Detection

ChangeFormer: A Transformer-Based Siamese Network for Change Detection (Under review at IGARSS-2022) Wele Gedara Chaminda Bandara, Vishal M. Patel Her

Wele Gedara Chaminda Bandara 214 Dec 29, 2022
Code for Low-Cost Algorithmic Recourse for Users With Uncertain Cost Functions

EMS-COLS-recourse Initial Code for Low-Cost Algorithmic Recourse for Users With Uncertain Cost Functions Folder structure: data folder contains raw an

Prateek Yadav 1 Nov 25, 2022
NeuralForecast is a Python library for time series forecasting with deep learning models

NeuralForecast is a Python library for time series forecasting with deep learning models. It includes benchmark datasets, data-loading utilities, evaluation functions, statistical tests, univariate m

Nixtla 1.1k Jan 03, 2023
PyTorch reimplementation of REALM and ORQA

PyTorch reimplementation of REALM and ORQA

Li-Huai (Allan) Lin 17 Aug 20, 2022
Customer-Transaction-Analysis - This analysis is based on a synthesised transaction dataset containing 3 months worth of transactions for 100 hypothetical customers.

Customer-Transaction-Analysis - This analysis is based on a synthesised transaction dataset containing 3 months worth of transactions for 100 hypothetical customers. It contains purchases, recurring

Ayodeji Yekeen 1 Jan 01, 2022
Official Repository for our ECCV2020 paper: Imbalanced Continual Learning with Partitioning Reservoir Sampling

Imbalanced Continual Learning with Partioning Reservoir Sampling This repository contains the official PyTorch implementation and the dataset for our

Chris Dongjoo Kim 40 Sep 18, 2022
MINERVA: An out-of-the-box GUI tool for offline deep reinforcement learning

MINERVA is an out-of-the-box GUI tool for offline deep reinforcement learning, designed for everyone including non-programmers to do reinforcement learning as a tool.

Takuma Seno 80 Nov 06, 2022
Gluon CV Toolkit

Gluon CV Toolkit | Installation | Documentation | Tutorials | GluonCV provides implementations of the state-of-the-art (SOTA) deep learning models in

Distributed (Deep) Machine Learning Community 5.4k Jan 06, 2023
GUI for a Vocal Remover that uses Deep Neural Networks.

GUI for a Vocal Remover that uses Deep Neural Networks.

4.4k Jan 07, 2023
Code implementing "Improving Deep Learning Interpretability by Saliency Guided Training"

Saliency Guided Training Code implementing "Improving Deep Learning Interpretability by Saliency Guided Training" by Aya Abdelsalam Ismail, Hector Cor

8 Sep 22, 2022
Normalizing Flows with a resampled base distribution

Resampling Base Distributions of Normalizing Flows Normalizing flows are a popular class of models for approximating probability distributions. Howeve

Vincent Stimper 24 Nov 03, 2022
Unified learning approach for egocentric hand gesture recognition and fingertip detection

Unified Gesture Recognition and Fingertip Detection A unified convolutional neural network (CNN) algorithm for both hand gesture recognition and finge

Mohammad 227 Dec 25, 2022
DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.

DeepSpeed+Megatron trained the world's most powerful language model: MT-530B DeepSpeed is hiring, come join us! DeepSpeed is a deep learning optimizat

Microsoft 8.4k Dec 28, 2022
PyTorch implementation of MLP-Mixer

PyTorch implementation of MLP-Mixer MLP-Mixer: an all-MLP architecture composed of alternate token-mixing and channel-mixing operations. The token-mix

Duo Li 33 Nov 27, 2022
Charsiu: A transformer-based phonetic aligner

Charsiu: A transformer-based phonetic aligner [arXiv] Note. This is a preview version. The aligner is under active development. New functions, new lan

jzhu 166 Dec 09, 2022
PartImageNet is a large, high-quality dataset with part segmentation annotations

PartImageNet: A Large, High-Quality Dataset of Parts We will release our dataset and scripts soon after cleaning and approval. Introduction PartImageN

Ju He 77 Nov 30, 2022
Unofficial PyTorch implementation of the Adaptive Convolution architecture for image style transfer

AdaConv Unofficial PyTorch implementation of the Adaptive Convolution architecture for image style transfer from "Adaptive Convolutions for Structure-

65 Dec 22, 2022
A Shading-Guided Generative Implicit Model for Shape-Accurate 3D-Aware Image Synthesis

A Shading-Guided Generative Implicit Model for Shape-Accurate 3D-Aware Image Synthesis Figure: Shape-Accurate 3D-Aware Image Synthesis. A Shading-Guid

Xingang Pan 115 Dec 18, 2022
O2O-Afford: Annotation-Free Large-Scale Object-Object Affordance Learning (CoRL 2021)

O2O-Afford: Annotation-Free Large-Scale Object-Object Affordance Learning Object-object Interaction Affordance Learning. For a given object-object int

Kaichun Mo 26 Nov 04, 2022
Meli Data Challenge 2021 - First Place Solution

My solution for the Meli Data Challenge 2021

Matias Moreyra 23 Mar 09, 2022