Windows>> neural-network-builder>> 返回
项目作者： arvindvs

项目描述 ：
Fully-connected neural network constructor and trainer for supervised learning
高级语言： Python
项目地址: git://github.com/arvindvs/neural-network-builder.git
创建时间： 2018-08-14T20:18:41Z
项目社区：https://github.com/arvindvs/neural-network-builder
开源协议：
下载


Neural Network Builder

Build and train a standard fully-connected neural network.
Choose the size of input, number of layers, type of activation function, and number of output classes in your network
Train your network by providing data inputs and inputs with a specific learning rate.
Use your fully trained network to perform classification or regression!

Getting Started

Prerequisites

Import NumPy

Usage

Creating architecture

Create a NeuralNetwork object by specifying the number of nodes in each layer as
an array (e.g. [256, 60, 10] for an array with input size of 256, hidden layer size of 60
and output size of 10). Specify the learning rate you want to use at when creating
the NeuralNetwork.

Training

Call the train function on the neural network using a set of data. The number of
inputs and outputs should match. Specify the number of epochs to be used in training.

Saving and Reading NN object

Functionality coming soon…