Processing math: 50%

alex_bn_lee

导航

< 2025年2月 >
26 27 28 29 30 31 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 1
2 3 4 5 6 7 8

统计

09 2020 档案

【486】机器学习的通用工作流程
摘要:参考:Python深度学习 P89 目录 定义问题,收集数据集 选择衡量成功的指标 确定评估方法 准备数据 开发比基准更好的模型 扩大模型规模:开发过拟合的模型 模型正则化与调节超参数 一、定义问题,收集数据集 首先,你必须定义所面对的问题。 你的输入数据是什么?你要预测什么? 你面对的是什么类型的 阅读全文

posted @ 2020-09-27 17:00 McDelfino 阅读(162) 评论(0) 推荐(0) 编辑

【485】deep learning 防止 overfitting 的方法
摘要:获取更多的训练数据 减小网络的容量(层数,神经元数量) 添加权重正则化(l1 和 l2) 添加 dropout 添加正则化来防止 overfitting 一种常见的降低 overfitting 的方法就是强制让模型权重只能取较小的值,从而限制模型的复杂度,这使得权重值的分布更加规则(regular) 阅读全文

posted @ 2020-09-27 11:39 McDelfino 阅读(269) 评论(0) 推荐(0) 编辑

【484】Book - Deep Learning with Python 相关说明与代码
摘要:目录: 一、相关函数说明 imdb.load_data() numpy array 可以通过 list 将所有索引赋值 善用 enumerate 函数 通过 numpy.array 来对比两个列表相似性 标签为整数,用 sparse_categorical_crossentropy 损失函数 num 阅读全文

posted @ 2020-09-25 21:14 McDelfino 阅读(220) 评论(0) 推荐(0) 编辑

【483】Keras 中 LSTM 与 BiLSTM 语法
摘要:参考:Keras-递归层Recurrent官方说明 参考:Keras-Bidirectional包装器官方说明 LSTM(units=32, input_shape=(10, 64)) units=32:输出神经元个数 input_shape=(10, 64):输入数据形状,10 代表时间序列的长度 阅读全文

posted @ 2020-09-24 23:55 McDelfino 阅读(2767) 评论(0) 推荐(0) 编辑

【482】Keras 实现 LSTM & BiLSTM
摘要:参考:Keras 实现 LSTM 参考:Keras-递归层Recurrent官方说明 参考:GitHub - Keras LSTM 参考:GitHub - Keras BiLSTM LSTM 是优秀的循环神经网络 (RNN) 结构,而 LSTM 在结构上也比较复杂,对 RNN 和 LSTM 还稍有疑 阅读全文

posted @ 2020-09-24 22:16 McDelfino 阅读(5483) 评论(1) 推荐(0) 编辑

【481】全网讲RNN与LSTM最好的资料
摘要:参考:李宏毅老师YouTube课程-ML Lecture 21-1: Recurrent Neural Network (Part I) 里面用鲜活的例子把复杂的关系剖析开了,通过不同的绘图模式更好理解。(2020年9月23日留) 参考:博客园引用的截图 阅读全文

posted @ 2020-09-23 18:44 McDelfino 阅读(238) 评论(0) 推荐(0) 编辑

【480】Keras实操
摘要:参考:Keras - Python Deep Learning Neural Network API Keras With TensorFlow Prerequisites - Getting Started With Neural Networks TensorFlow And Keras GPU 阅读全文

posted @ 2020-09-22 13:53 McDelfino 阅读(110) 评论(0) 推荐(0) 编辑

【479】cross-entropy与softmax的求导
摘要:cross-entropy 损失函数 L=ln(pc)=ln(softmax(yc)) p_c = softmax(y_c) = \frac{e^{y_c}}{e^{y_1} + ... + e^{y_c} + ... + e^{y_n}} 阅读全文

posted @ 2020-09-13 23:38 McDelfino 阅读(169) 评论(0) 推荐(0) 编辑

【478】Victor Zhou深度学习链接
摘要:An Introduction to Neural Networks Build your first neural network with Keras An Introduction to Convolutional Neural Networks Natural Language Proces 阅读全文

posted @ 2020-09-02 23:13 McDelfino 阅读(204) 评论(0) 推荐(0) 编辑

点击右上角即可分享
微信分享提示