Giter Site home page Giter Site logo

mmdetection-new's Introduction

mmdetection-new

一、mmdetection的安装(目前官方只支持Linux系统安装)

1,安装环境:

  Ubuntu16.04   CUDA9.0   CUDNN8.0   Pytorch1.1   Python3.7 2,这里建议使用单独的虚拟环境:

1,终端创建虚拟环境:conda create -n  虚拟环境名称  python=3.7

2,进入虚拟环境: source activate 虚拟环境名称 (退出指令:source deactivate)

3,利用conda 安装pytorch 和torchvision

conda install pytorch torchvision -c pytorch

4,下载mmdetection工具并配置虚拟环境:

 下载地址:https://github.com/python-bookworm/mmdetection-new

5,安装依赖库:

pip install -r requirements.txt

二、配置cascade_rcnn训练数据:

1,数据准备:

训练的原始样本图片放到mmdetection/data/coco/train2017文件夹下;用于验证的原始图片数据放data/coco/val2017

训练的标注数据(xml)放到mmdetection/data/coco/annotations/train文件夹;用于验证的标准数据(xml)放

mmdetection/data/coco/annotations/val文件夹;

2,xml转json:

打开mmdetection/xml2json.py,修改convert()中的文件路径,运行xml2json.py;

3,网络参数配置:cascade_rcnn_r50_fpn_1x.py

打开mmdetection/configs/cascade_rcnn_r50_fpn_1x.py:

     1,修改num_classes:#分类器的类别数量+1,+1是多了一个背景的类型;

     2,修改data_root:#数据集根目录

     3,修改img_scale:#输入的图像尺寸

     4,修改imgs_per_gpu:#每个gpu计算的图像数量

     5,修改workers_per_gpu:#每个gpu分配的线程数

     6,修改ann_file和img_prefix:#数据集annotation路径和数据集路径

     7,修改optimizer中的lr:#学习率,计算公式:imgs_per_gpu*0.00125

     8,修改total_epochs:#训练轮数

     9,work_dir:#log文件和模型文件存储路径

4,打开mmdetection/mmdet/datasets/coco.py:

     修改CLASSES中类别:#修改成自己的类别

三、运行train.py开始训练模型:

终端运行指令:python tools/train.py  configs/cascade_rcnn_r50_fpn_1x.py

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.