56 Star 150 Fork 53

时间坐标 / paradox

Create your Gitee Account
Explore and code with more than 6 million developers,Free private repositories !:)
Sign up
This repository doesn't specify license. Without author's permission, this code is only for learning and cannot be used for other purposes.
Clone or download
Cancel
Notice: Creating folder will generate an empty file .keep, because not support in Git
Loading...
README.md

Paradox 小型深度学习框架

Version Version Version

用python3和numpy实现一个简单的深度学习框架。

GraphExample

依赖

Name Version
Python 3.5.0+
Numpy 1.13.0+

开始

使用Paradox对进行梯度计算,并输出x的梯度。 整个过程由图计算和自动求导完成。

import paradox as pd

k = pd.Constant([[2, 3], [1, 1]], name='k')
b = pd.Constant([[7], [3]], name='b')
x = pd.Variable([[0], [0]], name='x')

y = k @ x + b

print(pd.Engine(y).gradient(x).value)

运行结果

[[ 3.]
 [ 4.]]

功能实现

  • 图计算(Graph Computing)。
  • 自动求导(Auto Gradient)。
  • 代数系统。
  • 梯度下降。
  • 神经网络API。
  • 卷积神经网络(Convolutional Neural Network)。

联系

Author ict
QQ 405340537
E-mail ictxiangxin@hotmail.com

Comments ( 6 )

Sign in for post a comment

About

纯python3和numpy实现的符号计算深度学习框架。 spread retract
Cancel

Releases

No release

Gitee Metrics

Contributors

All

Activities

load more
can not load any more
Python
1
https://git.oschina.net/ictxiangxin/paradox.git
git@git.oschina.net:ictxiangxin/paradox.git
ictxiangxin
paradox
paradox
master

Search

103611 48b8ff67 1899542 103622 4d02230c 1899542