keras
Lzj000lzj
这个作者很懒,什么都没留下…
展开
-
keras Sequential堆叠网络实现MLP多分类
数据准备import numpy as npimport pandas as pdfrom sklearn.datasets import load_iris from sklearn.model_selection import train_test_split#载入数据集 iris = load_iris()print(pd.DataFrame(iris.data).desc...原创 2019-06-28 15:05:07 · 1192 阅读 · 0 评论 -
keras 函数式api构建模型
函数式api构建模型input_x = tf.keras.Input(shape=(72,))hidden1 = layers.Dense(32, activation='relu')(input_x)hidden2 = layers.Dense(16, activation='relu')(hidden1)pred = layers.Dense(10, activation='softm...转载 2019-06-28 16:44:51 · 488 阅读 · 0 评论 -
keras.Sequential 实现简单的文本分类
IMDB数据imdb=keras.datasets.imdb(train_x, train_y), (test_x, text_y)=keras.datasets.imdb.load_data(num_words=10000)#参数num_words=10000表示数据集保留了最常出现的10,000个单词。为了保持数据大小的可处理性,罕见的单词会被丢弃。print(type(train_x...转载 2019-06-29 10:01:32 · 551 阅读 · 0 评论 -
keras处理欠拟合和过拟合
baselineimport tensorflow.keras.layers as layersbaseline_model = keras.Sequential([ layers.Dense(16, activation='relu', input_shape=(NUM_WORDS,)), layers.Dense(16, activation='relu'), l...转载 2019-06-29 10:58:15 · 1736 阅读 · 0 评论 -
keras使用word2cev在imdb数据集上做词嵌入,构建embedding矩阵
数据加载import numpy as npimport pandas as pdfrom sklearn.model_selection import train_test_splitfrom sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score, confusion_matrixf...转载 2019-07-10 18:28:55 · 1097 阅读 · 0 评论 -
keras之权重初始化
在神经网络训练中,好的权重 初始化会加速训练过程。下面说一下kernel_initializer 权重初始化的方法。不同的层可能使用不同的关键字来传递初始化方法,一般来说指定初始化方法的关键字是kernel_initializer 和 bias_initializermodel.add(Dense(64, kernel_initializer=initializers.random_norma...转载 2019-06-30 20:34:16 · 11474 阅读 · 0 评论 -
用keras使用glove预训练的词向量来构建实验的embedding矩阵-以Jigsaw Unintended Bias in Toxicity Classification比赛baseline为例
数据加载import numpy as np # linear algebraimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)import os# print(os.listdir("../input"))# print(os.listdir("../input/crawl300d2m"))...原创 2019-07-11 16:27:37 · 1452 阅读 · 1 评论