DenoisingAutoEncoder:堆叠式降噪自动编码器的Python实现,用于无监督学习高级特征表示-源码

上传者: 42102358 | 上传时间: 2021-08-20 16:02:55 | 文件大小: 16.55MB | 文件类型: ZIP
去噪自动编码器 可以训练去噪自动编码器,以无人监督的方式学习特征空间的高级表示。 可以通过将经过预训练的自动编码器的一层一层堆叠在一起来创建深度神经网络。 整个网络的培训分三个阶段进行: 1.预训练:在此阶段中,对每个层进行训练,以从损坏的版本中重建原始数据。 破坏输入的不同有效方法包括: -添加小高斯噪声-将变量随机设置为任意值-随机将输入变量设置为0 2.学习:在此阶段中,将S形层和softmax层放置在堆栈的顶部,并接受有关分类任务的培训。 3.微调:使用标准反向传播算法对整个网络进行微调 #创建堆叠降噪自动编码器的结构sDA = StackedDA([300,100]) # Pre-train layers one at a time, with 50% Salt and Pepper noise sDA.pre_train(X[:1000], rate=0.5, n_iters

文件下载

资源详情

[{"title":"( 38 个子文件 16.55MB ) DenoisingAutoEncoder:堆叠式降噪自动编码器的Python实现,用于无监督学习高级特征表示-源码","children":[{"title":"DenoisingAutoEncoder-master","children":[{"title":"Layers.pyc <span style='color:#111;'> 3.50KB </span>","children":null,"spread":false},{"title":"SDA_layers.py <span style='color:#111;'> 2.16KB </span>","children":null,"spread":false},{"title":"Layers.py <span style='color:#111;'> 2.86KB </span>","children":null,"spread":false},{"title":"results","children":[{"title":"dA_non_stochastic_regularized.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false},{"title":"dA_non_stochastic_regularized_90percent_dropouts.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false},{"title":"somDA_1000k_MSE_0.2_SP.png <span style='color:#111;'> 60.69KB </span>","children":null,"spread":false},{"title":"dA_non_stochastic_regularized_90percent_dropouts_9000_epochs.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false},{"title":"dA_non_stochastic_regularized_90percent_dropouts_4700_epochs_20_20.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false},{"title":"res_dA.png <span style='color:#111;'> 76.96KB </span>","children":null,"spread":false},{"title":"dA_non_stochastic_regularized_90percent_dropouts_4700_epochs.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false},{"title":"somDA_SP_500k.png <span style='color:#111;'> 62.83KB </span>","children":null,"spread":false},{"title":"Thumbs.db <span style='color:#111;'> 258.00KB </span>","children":null,"spread":false},{"title":"dA_non_stochastic.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false},{"title":"dA_stochastic_noisy_unit.png <span style='color:#111;'> 77.62KB </span>","children":null,"spread":false}],"spread":false},{"title":"demo.pyc <span style='color:#111;'> 1.08KB </span>","children":null,"spread":false},{"title":"tools.pyc <span style='color:#111;'> 1018B </span>","children":null,"spread":false},{"title":"utils.py <span style='color:#111;'> 1.33KB </span>","children":null,"spread":false},{"title":"utils.pyc <span style='color:#111;'> 2.20KB </span>","children":null,"spread":false},{"title":"Noise.pyc <span style='color:#111;'> 2.50KB </span>","children":null,"spread":false},{"title":"demo.py <span style='color:#111;'> 997B </span>","children":null,"spread":false},{"title":"README.md <span style='color:#111;'> 2.09KB </span>","children":null,"spread":false},{"title":"tools.py <span style='color:#111;'> 845B </span>","children":null,"spread":false},{"title":"mnist","children":[{"title":"mnist.pkl.gz <span style='color:#111;'> 15.42MB </span>","children":null,"spread":false}],"spread":true},{"title":"Noise.py <span style='color:#111;'> 1.09KB </span>","children":null,"spread":false},{"title":"Archive","children":[{"title":"lggd.py <span style='color:#111;'> 4.20KB </span>","children":null,"spread":false},{"title":"dA.py <span style='color:#111;'> 4.25KB </span>","children":null,"spread":false},{"title":"Loss.py <span style='color:#111;'> 906B </span>","children":null,"spread":false},{"title":"SDA.pyc <span style='color:#111;'> 1.61KB </span>","children":null,"spread":false},{"title":"somDA_gaussian_noise.png <span style='color:#111;'> 69.71KB </span>","children":null,"spread":false},{"title":"res_dA.png <span style='color:#111;'> 76.95KB </span>","children":null,"spread":false},{"title":"somDA_no_noise.png <span style='color:#111;'> 48.32KB </span>","children":null,"spread":false},{"title":"somDA_dropout_at_the_input.png <span style='color:#111;'> 52.62KB </span>","children":null,"spread":false},{"title":"SDA.py <span style='color:#111;'> 2.63KB </span>","children":null,"spread":false},{"title":"dA.pyc <span style='color:#111;'> 3.90KB </span>","children":null,"spread":false},{"title":"somDA_cross_entropy_no_hidden.png <span style='color:#111;'> 61.27KB </span>","children":null,"spread":false},{"title":"Loss.pyc <span style='color:#111;'> 1.51KB </span>","children":null,"spread":false},{"title":"somDA_0.5_hidden_noise_MSE.png <span style='color:#111;'> 61.93KB </span>","children":null,"spread":false}],"spread":false},{"title":"SDA_layers.pyc <span style='color:#111;'> 2.91KB </span>","children":null,"spread":false}],"spread":false}],"spread":true}]

评论信息

免责申明

【只为小站】的资源来自网友分享,仅供学习研究,请务必在下载后24小时内给予删除,不得用于其他任何用途,否则后果自负。基于互联网的特殊性,【只为小站】 无法对用户传输的作品、信息、内容的权属或合法性、合规性、真实性、科学性、完整权、有效性等进行实质审查;无论 【只为小站】 经营者是否已进行审查,用户均应自行承担因其传输的作品、信息、内容而可能或已经产生的侵权或权属纠纷等法律责任。
本站所有资源不代表本站的观点或立场,基于网友分享,根据中国法律《信息网络传播权保护条例》第二十二条之规定,若资源存在侵权或相关问题请联系本站客服人员,zhiweidada#qq.com,请把#换成@,本站将给予最大的支持与配合,做到及时反馈和处理。关于更多版权及免责申明参见 版权及免责申明