alex_bn_lee

导航

上一页 1 ··· 128 129 130 131 132 133 134 135 136 ··· 245 下一页

2020年6月20日 #

【471】Keras 深度神经网络实现

摘要: 参考:Keras 中文文档 参考:开始使用 Keras Sequential 顺序模型 Keras 的核心数据结构是 model,一种组织网络层的方式。最简单的模型是 Sequential 顺序模型,它由多个网络层线性堆叠。对于更复杂的结构,你应该使用 Keras 函数式 API,它允许构建任意的神 阅读全文

posted @ 2020-06-20 08:21 McDelfino 阅读(1352) 评论(0) 推荐(0)

2020年6月19日 #

【470】Keras API 参考说明

摘要: 参考:Keras API reference 参考:Keras layers API 目录: to_categorical() randint() random() Sequential() Model training APIs compile() fit() evaluate() predict 阅读全文

posted @ 2020-06-19 10:47 McDelfino 阅读(392) 评论(0) 推荐(0)

2020年6月15日 #

【469】RNN, LSTM参考博客

摘要: 参考:【推荐】ML Lecture 21-1: Recurrent Neural Network (Part I) 参考:Illustrated Guide to Recurrent Neural Networks 参考:Illustrated Guide to LSTM’s and GRU’s: 阅读全文

posted @ 2020-06-15 17:25 McDelfino 阅读(324) 评论(0) 推荐(0)

2020年5月29日 #

【468】numpy 数组获取对角线数值

摘要: 数组的索引也可以通过数组来添加 >>> a = np.arange(0, 25) >>> a.reshape(5, 5) array([[ 0, 1, 2, 3, 4], [ 5, 6, 7, 8, 9], [10, 11, 12, 13, 14], [15, 16, 17, 18, 19], [2 阅读全文

posted @ 2020-05-29 21:31 McDelfino 阅读(1935) 评论(0) 推荐(0)

【468】存储数据变量 — pickle 模块

摘要: 参考:python pickle 模块的使用 用于 python 中特有的类型数据存储。接触这个模块是看深度学习的时候,将训练后的权重参数通过 pickle 存储,然后之后再用的时候可以通过 pickle 在提取,很方便。主要是四个方法: pickle.dump(obj, file, protoco 阅读全文

posted @ 2020-05-29 14:24 McDelfino 阅读(340) 评论(0) 推荐(0)

上一页 1 ··· 128 129 130 131 132 133 134 135 136 ··· 245 下一页