Coder Social home page Coder Social logo

tfjs_test's Introduction

tensorflow: test-project

使用webcam识别。目前只有我一个人的样本集。

cd tfjs
npm install
npm run server

我的前端非常之烂_(:3」∠)_

demo:https://qxy65535.github.io/

目前能检测握拳和手掌,好像现在只有这个样子的识别是最稳的。。我好难过

最近工作

last updated: 2018/10/23

以下工作的python版本均为python 3.x

1、制作样本数据集

python camera.py

数据将生成于当前目录下的data文件夹下

  • 这个是根据我的摄像头写的。。拍摄224x224大小的文件。。我不知道大家的分辨率。。不保证没有BUG。。有需求可以自己写或者魔改。。

2、retrain mobilenet

训练方法:

参考https://www.tensorflow.org/hub/tutorials/image_retraining

下载retrain脚本

curl -LO https://github.com/tensorflow/hub/raw/master/examples/image_retraining/retrain.py

retrain自定义数据集

python retrain.py --image_dir train \
  --tfhub_module https://tfhub.dev/google/imagenet/mobilenet_v1_025_224/feature_vector/1 \
  --how_many_training_steps=your_steps \
  --validation_batch_size=your_batch_size

训练后结果默认放置在/tmp中,可以自定义地设置更多选项。

3、制作tensorflow.js可用模型

将得到的tensorflow freeze模型转化为tensorflow.js可加载的模型

tensorflowjs_converter \
  --input_format=tf_frozen_model \
  --output_node_names='final_result' \
  --saved_model_tags=serve \
  /tmp/output_graph.pb \
  ~/mobilenet

retrain的原理是在mobilenet的最后延接一层并命名为 final_result(可在retrain.py时自定义),然后用原mobilenet网络的输出作为其输入,仅训练最后一层(迁移学习)。本质上同吃豆人demo的做法,原理可参考吃豆人源码https://github.com/tensorflow/tfjs-examples/tree/master/webcam-transfer-learning。

我也有用keras复现吃豆人那种网络。。截取前部分mobilenet然后续接两层全连接。。然而分类准确率一直上不去。。结果也有(fei)点(chang)惨淡就先不放脏脏的源码了_(:3」∠)_

  • 我觉得会不会是在我的数据集上过拟合了导致的?? 但转念一下它连我都认不出来好像不应该啊

4、使用tensorflow.js和webcam识别并输出

详见本项目tfjs下的源码

前端水平极差。。抄袭吃豆人

坑点

1、使用tensorflowjs_converter的时候,在windows下获得的文件无论如何不能被载入。。会报错。。比如

uncaught (in promise) error: based on the provided shape, [3,3,8,1], and dtype float32, the tensor should have 72 values but has 3

其实也不是完全不能载入啦谜一样的就成功过一次。。有人说是生产环境导致的。。搞得我删了又建建了又删parcel和webpack换来换去。。

然而在ubuntu下执行它得到的文件完全没出过问题啊!!我好生气!!

2、好像暂时没有

tfjs_test's People

Contributors

qxy65535 avatar

Watchers

James Cloos avatar  avatar

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.