Giter Club home page Giter Club logo

quiz-w8-doc's Introduction

第9周作业描述

文件目录结构

├── create_quiz_w8_tf_record.sh
├── data
│   ├── labels_items.txt
│   ├── pet_train.record
│   ├── pet_val.record
│   ├── pre_trained
│   ├── quiz-w8-data│   
│   ├── ssd_mobilenet_v1_pets.config
│   └── test.jpg
├── inference.py
├── object_detection
├── output
│   ├── eval
│   ├── exported_graphs
│   ├── output.png
│   ├── ssd_mobilenet_v1_pets.config
│   └── train
├── readme.md
├── run_local.py
├── run_local.sh
├── run.py
├── run.sh
└── slim    

说明

  • tensorflow工具包下载地址 https://github.com/tensorflow/models/tree/r1.5,
  • 一定要切换到r1.5分支去下载, r1.5分支是基于tensorflow1.5.0写的, r1.6分支是基于tensorflow1.6写的,以此类推.
  • 由于作业只涉及object_detection,slim这俩个工具,所以只将这俩个拷贝到项目下.
  • 使用pycocotool3.5.1编译 object_detection/protos 下的proto文件

output目录

  • output目录为本地运行结果
  • 本地运行是通过在tinymind训练好后,下载tinymind上/output/train下的最新训练文件到本地的train目录,并修改checkpoint文件后
  • 再在本地运行export_inference_graph.py 及 inference.py 输出需要的结果

数据集

  • 下载https://gitee.com/ai100/quiz-w8-data.git数据集,
  • cp object_detection/dataset_tools/create_pet_tf_record.py object_detection/dataset_tools/create_quiz_tf_record.py
  • 编辑create_quiz_tf_record.py 文件使其适应quiz-w8数据集
  • 运行下面脚步得到tf.record数据
     sh create_quiz_w8_tf_record.sh
    

预训练模型

配置文件修改

  • ssd_mobilenet_v1_pets.config的5个路径修改
  • eval_config.num_examples修改
  • model.ssd.num_classes修改

模型训练

运行脚本

python run.py

train_log

结果评估

output

心得体会

  • 本次作业主要是熟悉了使用tensorflow工具包去对特定数据集的目标检测
  • 对tensorflow工具包的版本问题更熟悉了
  • 熟悉了tf.record数据集的生成
  • 熟悉对配置文件的修改,及可能出现的配置出错引起的报错.

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.