rbm 对比散度cd

weixin_42673583

已于 2022-07-29 17:26:35 修改

阅读量164

点赞数

分类专栏： matlab 机器学习文章标签： matlab

于 2022-07-23 11:55:37 首次发布

本文链接：https://blog.csdn.net/weixin_42673583/article/details/125945822

版权

%% training processe
for i = 1:max_itera
    i
errvec=[];                                      % initialize error history for every iteration
ordering = randperm(size(data,1),Options.Sz_mb);% randomly shose a batch of data
mini_batch = data(ordering, :);                 % load our mini-Batch
    
    for j = 1:N_gs % start gibbs sampling using energy function
                                            %mini_batch 就是可见层 h
        hidden_p =  sigmoid(mini_batch * W + repmat(h_bias,size(mini_batch,1),1));         %P(h|v) Find hidden units by sampling the visible layer.                                           
        visible_p =sigmoid( hidden_p* W' +repmat(v_bias,size(mini_batch,1),1));   %P(v|h) Find visible units by sampling from the hidden ones.                                           
        bP = sigmoid(visible_p * W + repmat(h_bias,size(mini_batch,1),1));       % 下一轮的隐藏层 Last step : Find