wehub-resource-sync
127 lines
6.0 KiB
Markdown
127 lines
6.0 KiB
Markdown
<!-- WEHUB_ZH_README -->
|
||
> [!NOTE]
|
||
> 本文档由 WeHub 基于上游 README 翻译整理，属于社区翻译，非官方中文文档。
|
||
> [English](./README.en.md) · [原始项目](https://github.com/karpathy/convnetjs) · [上游 README](https://github.com/karpathy/convnetjs/blob/HEAD/Readme.md)
|
||
> 原作者、版权与许可证归属以原始项目及本仓库 LICENSE 文件为准。
|
||
|
||
# ConvNetJS
|
||
|
||
ConvNetJS 是神经网络的 JavaScript 实现，并附带精美的基于浏览器的演示。目前支持：
|
||
|
||
- 常见的**神经网络模块**（全连接层、非线性激活）
|
||
- 分类（SVM/Softmax）与回归（L2）**代价函数**（cost functions）
|
||
- 可指定并训练处理图像的**卷积网络**（Convolutional Networks）
|
||
- 基于 Deep Q Learning 的实验性**强化学习**（Reinforcement Learning）模块
|
||
|
||
更多详细信息请参阅主站 [convnetjs.com](http://convnetjs.com)
|
||
|
||
**注意**：我已不再积极维护 ConvNetJS，因为我实在没有时间。我认为此时 npm 仓库可能已无法正常工作。
|
||
|
||
## 在线演示
|
||
- [MNIST 手写数字卷积神经网络](http://cs.stanford.edu/~karpathy/convnetjs/demo/mnist.html)
|
||
- [CIFAR-10 卷积神经网络](http://cs.stanford.edu/~karpathy/convnetjs/demo/cifar10.html)
|
||
- [玩具 2D 数据](http://cs.stanford.edu/~karpathy/convnetjs/demo/classify2d.html)
|
||
- [玩具 1D 回归](http://cs.stanford.edu/~karpathy/convnetjs/demo/regression.html)
|
||
- [在 MNIST 手写数字上训练自编码器](http://cs.stanford.edu/~karpathy/convnetjs/demo/autoencoder.html)
|
||
- [Deep Q Learning 强化学习演示](http://cs.stanford.edu/people/karpathy/convnetjs/demo/rldemo.html)
|
||
- [图像回归（"绘画"）](http://cs.stanford.edu/~karpathy/convnetjs/demo/image_regression.html)
|
||
- [MNIST 上 SGD/Adagrad/Adadelta 对比](http://cs.stanford.edu/people/karpathy/convnetjs/demo/trainers.html)
|
||
|
||
## 示例代码
|
||
|
||
下面是一个定义**双层神经网络**并在单个数据点上进行训练的最小示例：
|
||
|
||
```javascript
|
||
// species a 2-layer neural network with one hidden layer of 20 neurons
|
||
var layer_defs = [];
|
||
// input layer declares size of input. here: 2-D data
|
||
// ConvNetJS works on 3-Dimensional volumes (sx, sy, depth), but if you're not dealing with images
|
||
// then the first two dimensions (sx, sy) will always be kept at size 1
|
||
layer_defs.push({type:'input', out_sx:1, out_sy:1, out_depth:2});
|
||
// declare 20 neurons, followed by ReLU (rectified linear unit non-linearity)
|
||
layer_defs.push({type:'fc', num_neurons:20, activation:'relu'});
|
||
// declare the linear classifier on top of the previous hidden layer
|
||
layer_defs.push({type:'softmax', num_classes:10});
|
||
|
||
var net = new convnetjs.Net();
|
||
net.makeLayers(layer_defs);
|
||
|
||
// forward a random data point through the network
|
||
var x = new convnetjs.Vol([0.3, -0.5]);
|
||
var prob = net.forward(x);
|
||
|
||
// prob is a Vol. Vols have a field .w that stores the raw data, and .dw that stores gradients
|
||
console.log('probability that x is class 0: ' + prob.w[0]); // prints 0.50101
|
||
|
||
var trainer = new convnetjs.SGDTrainer(net, {learning_rate:0.01, l2_decay:0.001});
|
||
trainer.train(x, 0); // train the network, specifying that x is class zero
|
||
|
||
var prob2 = net.forward(x);
|
||
console.log('probability that x is class 0: ' + prob2.w[0]);
|
||
// now prints 0.50374, slightly higher than previous 0.50101: the networks
|
||
// weights have been adjusted by the Trainer to give a higher probability to
|
||
// the class we trained the network with (zero)
|
||
```
|
||
|
||
如果你希望对图像进行预测，这里是一个小型**卷积神经网络**示例：
|
||
|
||
```javascript
|
||
var layer_defs = [];
|
||
layer_defs.push({type:'input', out_sx:32, out_sy:32, out_depth:3}); // declare size of input
|
||
// output Vol is of size 32x32x3 here
|
||
layer_defs.push({type:'conv', sx:5, filters:16, stride:1, pad:2, activation:'relu'});
|
||
// the layer will perform convolution with 16 kernels, each of size 5x5.
|
||
// the input will be padded with 2 pixels on all sides to make the output Vol of the same size
|
||
// output Vol will thus be 32x32x16 at this point
|
||
layer_defs.push({type:'pool', sx:2, stride:2});
|
||
// output Vol is of size 16x16x16 here
|
||
layer_defs.push({type:'conv', sx:5, filters:20, stride:1, pad:2, activation:'relu'});
|
||
// output Vol is of size 16x16x20 here
|
||
layer_defs.push({type:'pool', sx:2, stride:2});
|
||
// output Vol is of size 8x8x20 here
|
||
layer_defs.push({type:'conv', sx:5, filters:20, stride:1, pad:2, activation:'relu'});
|
||
// output Vol is of size 8x8x20 here
|
||
layer_defs.push({type:'pool', sx:2, stride:2});
|
||
// output Vol is of size 4x4x20 here
|
||
layer_defs.push({type:'softmax', num_classes:10});
|
||
// output Vol is of size 1x1x10 here
|
||
|
||
net = new convnetjs.Net();
|
||
net.makeLayers(layer_defs);
|
||
|
||
// helpful utility for converting images into Vols is included
|
||
var x = convnetjs.img_to_vol(document.getElementById('some_image'))
|
||
var output_probabilities_vol = net.forward(x)
|
||
```
|
||
|
||
## 入门
|
||
|
||
主站提供了[入门](http://cs.stanford.edu/people/karpathy/convnetjs/started.html) 教程。
|
||
|
||
完整的[文档](http://cs.stanford.edu/people/karpathy/convnetjs/docs.html) 也可在那里找到。
|
||
|
||
请参阅本项目的 **releases** 页面以获取压缩编译后的库；下方也提供了直接链接以方便使用（但请自行托管副本）
|
||
|
||
- [convnet.js](http://cs.stanford.edu/people/karpathy/convnetjs/build/convnet.js)
|
||
- [convnet-min.js](http://cs.stanford.edu/people/karpathy/convnetjs/build/convnet-min.js)
|
||
|
||
## 从 src/ 编译库到 build/
|
||
|
||
如果你想为库添加功能，需要修改 `src/` 中的代码，然后将库编译到 `build/` 目录。编译脚本会拼接 `src/` 中的文件，然后压缩结果。
|
||
|
||
编译通过 ant 任务完成：它会拼接 `src/` 中的源文件来编译 `build/convnet.js`，然后将结果压缩到 `build/convnet-min.js`。请确保已安装 **ant**（在 Ubuntu 上可简单地通过 *sudo apt-get install* 安装），然后 cd 进入 `compile/` 目录并运行：
|
||
|
||
$ ant -lib yuicompressor-2.4.8.jar -f build.xml
|
||
|
||
输出文件将位于 `build/`
|
||
## 在 Node 中使用
|
||
|
||
该库也可在 *node.js* 中使用：
|
||
|
||
1. 安装：`$ npm install convnetjs`
|
||
2. 使用：`var convnetjs = require("convnetjs");`
|
||
|
||
## 许可证
|
||
|
||
MIT
|