最简单的python机器学习工具包。
ztlearn的Python项目详细描述
zeta learn
----
zeta learn是一个极简的python机器学习库,旨在提供快速、简单的模型原型设计。
zeta learn旨在通过使用简单的算法和容易的方法,提供对机器学习的广泛理解。实现的示例使
成为研究人员和学生的有用资源。
**gt;=2.0.0
features
——
——用于构建模型的类似keras的顺序api。
-基于numpy和matplotlib构建。
-带有易于实现的机器学习模型的示例文件夹。
install
——
-pip installztlearn
Examples
--------
Principal Component Analysis (PCA)
##################################
`DIGITS Dataset - PCA <https://github.com/jefkine/zeta-learn/blob/master/examples/digits/digits_pca.py>`_
=====================
.. 图片::examples/plots/results/pca/digits-pca.png
:align:center
:alt:digits-pca
:align:center
:alt:mnist pca
`k-means clustering(4个集群)<;https://github.com/jefkine/zeta learn/blob/master/examples/clusters/kmeans嫒clustering.py>;`
==
==
image:: /examples/plots/results/kmeans/k_means_4_clusters.png
:align: center
:alt: k-means (4 clusters)
Convolutional Neural Network (CNN)
##################################
`DIGITS Dataset Model Summary <https://github.com/jefkine/zeta learn/blob/master/examples/digits/digits\u cnn.py>;` ` ` `
===========
。代码::html
8)激活:relu0(32,8,8)brdropout(32,8,8)brbatchnormalization4096(32,8,8)brconv2d18496(64,8,8)br激活:relu(64,8,8)
最大池2dΩ0Ω(64,7,7)
辍学Ω0Ω(64,7,7)
批量规范化Ω6272Ω(64,7,7)
扁平化Ω0Ω(3136,)
致密化Ω803072Ω(256,)活化:relu0(256,brdropout0(256,brbatchnormalization512(256,brdense2570(10,)?
+————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————图片://examples/plots/results/cnn/digits_cnn_tiled_results.png
:align:center
:alt:digits cnn results tiled
>digits dataset model loss
=图片://examples/plots/results/cnn/digits_cnn_loss_graph.png
:align:center
:alt:digits model loss
digits dataset model accurity
======
。图片::/examples/plots/results/cnn/digits_cnn_accorrecity_graph.png
:align:center
:alt:digits model accorrecity
`mnist dataset model summary<;https://github.com/jefkine/zeta learn/blob/master/examples/mnist/mnist_cnn.py>;`<
=================
代码::html
mnist cnn
28,28)×br/>激活:relu×0×(32,28,28)×br/>退出×0×(32,28,28)×br/>批量规范化×50176×(32,28,28)×br/>conv2d×18496×(64,28,28)×br/>激活:relu×。0Ω(64,28,28)
最大池2dΩ(64,27,27)
辍学Ω0Ω(64,27,27)
批量规范化Ω93312Ω(64,27,27)
扁平化Ω0Ω(46656,)
致密?11944192?激活:relu?0?256,?br/>?dropout?0?256,?br/>?batchnormalization?512?256,?br/>?dense?2570?10,)?
+—————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————————图片::/examples/plots/results/cnn/mnist-cnn-cnn-tiled_results.png
:align:center
:alt:mnist-cnn结果tiled
>回归
`线性回归<;https://github.com/jefkine/zeta-learn/blob/master/examples/boston/boston/cnn-cnn-cnn结果tiled
>回归回归
波士顿线性回归.py>;`
==
==
…图片::examples/plots/results/regression/linear廑regression.png
:align:center
:alt:linear regression
`多项式回归<;https://github.com/jefkine/zeta learn/blob/master/examples/boston/boston廑polymonial廑regression.py>;`廑
==图片::examples/plots/results/regression/polymonial_regression.png
:align:center
:alt:polymonial regression
`elastic regression<;https://github.com/jefkine/zeta learn/blob/master/examples/boston/boston_elastic_regression.py>;` `<
===br/>。图片::/examples/plots/results/regression/elastic_regression.png
:对齐:居中
:alt:弹性回归