https://haillk.github.io/categories/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/ Recent content in 机器学习 on Haillk-Blog Hugo -- gohugo.io zh-CN Sat, 11 Apr 2020 01:16:17 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E6%B1%A0%E5%8C%96/ Sat, 11 Apr 2020 01:16:17 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E6%B1%A0%E5%8C%96/ <p>前进！前进！！不择手段的前进！</p> https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/AttentionGAN/ Sun, 29 Mar 2020 00:18:57 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/AttentionGAN/ <p>*</p> https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/SPA-GAN%E7%BD%91%E7%BB%9C%E6%9E%B6%E6%9E%84/ Mon, 23 Mar 2020 19:39:57 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/SPA-GAN%E7%BD%91%E7%BB%9C%E6%9E%B6%E6%9E%84/ * 论文内容： https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/SAGAN%E7%94%9F%E6%88%90%E5%99%A8%E5%92%8C%E5%88%A4%E5%88%AB%E5%99%A8/ Mon, 23 Mar 2020 15:30:18 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/SAGAN%E7%94%9F%E6%88%90%E5%99%A8%E5%92%8C%E5%88%A4%E5%88%AB%E5%99%A8/ * SAGAN ​ 生成器： 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 #将一些层封装成块，，注意这里有SN的参与 def block(input,output_channel,training=false): x = BatchNormalization(input) x = RELU() convtr = conv2dT(output_channels,3,2,padding=&#39;same&#39;) x = SN(convtr) x = BatchNormalization()&lt;------------------------------------------- x = RELU() convtr = https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/SA_NET/ Sun, 22 Mar 2020 18:14:38 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/SA_NET/ * 论文内容： 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 使用self-attention的SANET实现风格迁移 基本概念： ​ 内容图像lc https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E5%8D%B7%E7%A7%AFconv2d/ Fri, 20 Mar 2020 23:18:48 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E5%8D%B7%E7%A7%AFconv2d/ Conv2d conv2d参考,看这个足够了 参考原文 函数定义： 1 2 keras.layers.Conv2D(filters, kernel_size, strides=(1, 1), padding=&#39;valid&#39;, data_format=None, dilation_rate=(1, 1), activation=None, use_bias=True, kernel_initializer=&#39;glorot_uniform&#39;, bias_initializer=&#39;zeros&#39;, kernel_regularizer=None, bias_regularizer=None, activity_regularizer=None, kernel_constraint=None, bias_constraint=None) 该层创建一个卷积核，对层输入惊醒卷积以生成张量。如果user_bias为 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/tfrecord/ Mon, 16 Mar 2020 22:59:58 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/tfrecord/ TFrecord文件定义： 这是一种将图像数据和标签放在一起的二进制文件，能更好的利用内存，在tensorflow中快速的复制，移动，读取，存储 等等.. 本质上和x https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/tf2_%E7%94%A8tfdata%E5%8A%A0%E8%BD%BD%E6%95%B0%E6%8D%AE/ Mon, 16 Mar 2020 17:02:36 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/tf2_%E7%94%A8tfdata%E5%8A%A0%E8%BD%BD%E6%95%B0%E6%8D%AE/ <p>使用tf.data进行图像预处理</p> https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/DCGAN_tf2/ Sun, 15 Mar 2020 15:30:34 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/DCGAN_tf2/ 跟着tf官网过一遍DCGAN的算法流程，源地址 导入包 1 from __future__ import absolute_import, division, print_function, unicode_literals absolute_import模块：一般情况下，程序默认是先在系统的默认目录中寻找导入的包， https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E6%99%AE%E5%BD%92%E4%B8%80%E5%8C%96/ Sat, 07 Mar 2020 03:14:05 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E6%99%AE%E5%BD%92%E4%B8%80%E5%8C%96/ https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/GAN%E5%8F%91%E5%B1%95%E6%97%B6%E9%97%B4%E8%BD%B4/ Sat, 29 Feb 2020 17:10:06 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/GAN%E5%8F%91%E5%B1%95%E6%97%B6%E9%97%B4%E8%BD%B4/ https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E7%BA%BF%E6%80%A7%E4%BB%A3%E6%95%B0%E5%9F%BA%E6%9C%AC%E6%A6%82%E5%BF%B5/ Sat, 15 Feb 2020 20:22:12 +0800 https://haillk.github.io/post/%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0/%E7%BA%BF%E6%80%A7%E4%BB%A3%E6%95%B0%E5%9F%BA%E6%9C%AC%E6%A6%82%E5%BF%B5/