classicle

所属分类:后台框架
开发工具:Jupyter Notebook
文件大小:3977KB
下载次数:0
上传日期:2020-05-02 12:01:52
上 传 者sh-1993
说明:  NLP项目将文章分为五类:商业、娱乐、***、体育、科技
(NLP project to classify articles out of five classes: business, entertainment, politics, sport, tech)

文件列表:
LICENSE (1076, 2020-05-02)
__pycache__ (0, 2020-05-02)
__pycache__\classicle_model.cpython-37.pyc (1103, 2020-05-02)
__pycache__\testclassicle.cpython-37.pyc (1656, 2020-05-02)
classicle_model.py (1117, 2020-05-02)
classicle_v0.1.0.ipynb (73082, 2020-05-02)
classifybutton (0, 2020-05-02)
classifybutton\classifybutton (0, 2020-05-02)
classifybutton\classifybutton\__init__.py (0, 2020-05-02)
classifybutton\classifybutton\__pycache__ (0, 2020-05-02)
classifybutton\classifybutton\__pycache__\__init__.cpython-37.pyc (196, 2020-05-02)
classifybutton\classifybutton\__pycache__\classicle_model.cpython-37.pyc (1123, 2020-05-02)
classifybutton\classifybutton\__pycache__\settings.cpython-37.pyc (2335, 2020-05-02)
classifybutton\classifybutton\__pycache__\urls.cpython-37.pyc (1091, 2020-05-02)
classifybutton\classifybutton\__pycache__\views.cpython-37.pyc (1001, 2020-05-02)
classifybutton\classifybutton\__pycache__\wsgi.cpython-37.pyc (613, 2020-05-02)
classifybutton\classifybutton\asgi.py (405, 2020-05-02)
classifybutton\classifybutton\classicle_model.py (1106, 2020-05-02)
classifybutton\classifybutton\classifyclick.py (214, 2020-05-02)
classifybutton\classifybutton\settings.py (3123, 2020-05-02)
classifybutton\classifybutton\urls.py (850, 2020-05-02)
classifybutton\classifybutton\views.py (652, 2020-05-02)
classifybutton\classifybutton\wsgi.py (405, 2020-05-02)
classifybutton\db.sqlite3 (131072, 2020-05-02)
classifybutton\manage.py (634, 2020-05-02)
classifybutton\templates (0, 2020-05-02)
classifybutton\templates\classicle-ui.html (1210, 2020-05-02)
classifyclick.py (214, 2020-05-02)
data (0, 2020-05-02)
data\bbc-text.csv (5057493, 2020-05-02)
environment.yml (1845, 2020-05-02)
label_word_index.pickle (79, 2020-05-02)
model-bilstm.h5 (1973112, 2020-05-02)
requirements.yml (2442, 2020-05-02)
testclassicle.py (18873, 2020-05-02)
tokenizer.pickle (1092233, 2020-05-02)

[![License](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE) # Class*icle* :book: Class*icle* is a Natural Language Processing project that can classify your article out of five classes: business, entertainment, politics, sport, tech. It is read like _classical_ and it is a portmanteau word -- a blend of **class**ify and art**icle**. ## Demo Class*icle* demo is now on my [youtube channel](https://www.youtube.com/watch?v=qkPFRTR8QfQ) ![classicle_demo](https://user-images.githubusercontent.com/118***152/71850472-08b1d800-30dd-11ea-9***d-aea73a131f95.gif) ## How to use class*icle* 1. Clone this project and cd into the classifybutton folder: ``` $ git clone https://github.com/EzzEddin/classicle.git $ cd classicle $ cd classifybutton ``` 2. Using python3, run the following: ``` $ python3 manage.py runserver 127.0.0.1:8002 ``` 3. Go to the browser and write the server link, in the url, that you specified in the last command `127.0.0.1:8002` 4. The page you saw at the demo will apear so you can put your article in there to classify. ### Setup Install dependencies: - django - tensorflow - keras - pickle - numpy - csv - wget (optional) ## How class*icle* works Class*icle* is a project run on a django server just by clicking on *classify* button, it will run the python script which has the deep learning model. ## Data The data I used for training and testing is bbc-news articles available [here](http://mlg.ucd.ie/datasets/bbc.html). ## Acknowledgements The deployment and deep learning model in this project are inspired by [Browser-based Models with TensorFlow.js](https://www.coursera.org/learn/browser-based-models-tensorflow/home/welcome) course and [Natural Language Processing in TensorFlow](https://www.coursera.org/learn/natural-language-processing-tensorflow/home/welcome) course while running the model written in python script by a click on a button at django server is inspired by this [blog](https://www.hackanons.com/2019/04/run-python-script-on-clicking-html.html). I also want to thank my friend Nour for answering my questions whether in development or testing. Thank you, [guru](https://github.com/noureddin). ## License MIT, Copyright (C) 2020 by Ezz El Din Abdullah

近期下载者

相关文件


收藏者