ml-lab/auxiliary-deep-generative-models

语言: Python

git: https://github.com/ml-lab/auxiliary-deep-generative-models

半监督学习辅助深度生成模型的实现
Implementation of auxiliary deep generative models for semi-supervised learning
README.md (中文)

辅助深度生成模型

该存储库是研讨会上提出的辅助深度生成模型的实现 近似贝叶斯推断的进展<http://approximateinference.org>,NIPS 2015 文章<http://approximateinference.org/accepted/MaaloeEtAl2015.pdf>展示MNIST上最先进的表现和 将以扩展格式提交给ICML 2016,其中包含更多数据集。

实施基于Parmesan <https://github.com/casperkaae/parmesan>,Lasagne <http://github.com/Lasagne/Lasagne>和Theano <https://github.com/Theano/Theano> _图书馆。

安装

在执行python脚本之前,请确保已安装了这些要求。

安装

.. code-block :: bash

git clone https://github.com/casperkaae/parmesan.git   cd帕尔马干酪   python setup.py开发   pip安装numpy   pip安装seaborn   pip安装matplotlib   点击安装https://github.com/Theano/Theano/archive/master.zip   点击安装https://github.com/Lasagne/Lasagne/archive/master.zip

例子

存储库主要包括

  • 脚本在MNIST数据集上运行新模型,只有100个标签 - run_adgmssl_mnist.py。
  • 脚本评估训练模型(参见输出中的模型细节/。) - run_adgmssl_evaluation.py *。
  • iPython笔记本,所有培训都在一个scipt中实现 - run_adgmssl_mnist_notebook.ipynb。

有关更多详细信息,请参阅源代码和代码示例。

比较adgmssl,adgmssl与确定性辅助单元和dgmssl之间的训练收敛性。

.. image :: /output/train.png

分别显示辅助单元和潜在单元a和z的信息贡献。

.. image :: /output/diff.png

来自潜在空间的随机样本贯穿生成模型。

.. image :: /output/mnist.png

本文使用googletrans自动翻译,仅供参考, 原文来自github.com

en_README.md

Auxiliary Deep Generatives Models

This repository is the implementation of the auxiliary deep generative model presented at the workshop on
advances in approximate Bayesian inference <http://approximateinference.org>, NIPS 2015. The
article <http://approximateinference.org/accepted/MaaloeEtAl2015.pdf>
show state-of-the-art performance on MNIST and
will be submitted to ICML 2016 in an extended format where more datasets are included.

The implementation is build on the Parmesan <https://github.com/casperkaae/parmesan>, Lasagne <http://github.com/Lasagne/Lasagne> and Theano <https://github.com/Theano/Theano>_ libraries.

Installation

Please make sure you have installed the requirements before executing the python scripts.

Install

.. code-block:: bash

git clone https://github.com/casperkaae/parmesan.git
cd parmesan
python setup.py develop
pip install numpy
pip install seaborn
pip install matplotlib
pip install https://github.com/Theano/Theano/archive/master.zip
pip install https://github.com/Lasagne/Lasagne/archive/master.zip

Examples

The repository primarily includes

  • script running a new model on the MNIST datasets with only 100 labels - run_adgmssl_mnist.py.
  • script evaluating a trained model (see model specifics in output/.) - run_adgmssl_evaluation.py*.
  • iPython notebook where all training is implemented in a single scipt - run_adgmssl_mnist_notebook.ipynb.

Please see the source code and code examples for further details.

Comparison of the training convergence between the adgmssl, adgmssl with deterministic auxiliary units and the dgmssl.

.. image:: /output/train.png

Showing the information contribution from the auxiliary and the latent units a and z respecively.

.. image:: /output/diff.png

A random sample from the latent space run through the generative model.

.. image:: /output/mnist.png