qq_27047075的博客

Word2Vec模型中，主要有Skip-Gram和CBOW两种模型，从直观上理解，Skip-Gram是给定input word来预测上下文。而CBOW是给定上下文，来预测input word。

Tensorflow简单二分类任务，自定义训练循环

运行这个代码要安装一些包numpy == 1.17opencv-python >= 4.1torch >= 1.5matplotlibpycocotoolstqdmpillowtensorboard >= 1.141数据准备1.1 打标签使用labelimg 进行标签的生成，图片不要有中文路径和名称，打完标签时xml文件主要是执行这个软件的命令 python labelimg.py2 将图片放到JEPGimg文件夹下，images里也要一份3 将.

# coding: utf-8# In[1]:import tensorflow as tffrom tensorflow import kerasfrom tensorflow.keras import layers,Sequentialfrom tensorflow.examples.tutorials.mnist import input_dataimport matp...

# coding: utf-8# In[54]:import tensorflow as tffrom tensorflow import kerasfrom tensorflow.keras import layers,Sequentialtf.__version__# In[55]:#第一层3个神经元#第二层两个神经元network=Sequential([...

import tensorflow as tfimport numpy as npdef relu(temp): return tf.maximum(temp,0)def softmax2(z): exp=tf.exp(z) sum=tf.reduce_sum(tf.exp(z),axis=1,keepdims=True)#在第二个维度上求和，并且保持维度不变 ...

#感知机是二分类的线性模型，其输入是实例的特征向量，输出的是事例的类别，分别是+1和-1，属于判别模型。import numpy as npimport matplotlib.pyplot as pltc0=np.random.normal(6,1,100)c1=np.random.normal(3,1,100)c2=np.random.normal(3...

1.导入测试数据from sklearn.datasets import load_digitsimport numpy as npimport matplotlib.pyplot as pltdigits = load_digits()print(digits.data.shape)print(digits.target.shape)print(digits.images.sh...

#生成四个簇import numpy as npimport matplotlib.pyplot as pltc0x=np.random.normal(2,1,100).reshape(100,1)c0y=np.random.normal(2,1,100).reshape(100,1)c1x=np.random.normal(-2,1,100).reshape(100,1)c1y=n...

from sklearn.cluster import KMeanskmeans = KMeans(n_clusters = 3)kmeans.fit(x)y_kmeans = kmeans.predict(x)plt.scatter(x[:,0], x[:,1], c = y_kmeans, cmap='Dark2', s=50, alpha=0.5, marker='x')cen...

数据可能是高纬度的，资料标准化后会使得每个特征中的数值平均变为0（将每个特征的值都减掉原始资料中该特征的平均），标准差变为1StandardScaler计算训练集的平均值和标准差，以便测试数据及使用相同的变换。　　变换后各维特征有0均值，单位方差，也叫z-score规范化（零均值规范化），计算方式是将特征值减去均值，除以标准差。所有数减去其平均值的平方和，所得结果除以该组数...

1 梯度下降import numpy as npimport matplotlib.pyplot as pltx=np.array(range(0,50))np.random.seed(1)y=6*xlabel_y=y+np.random.random(50)*10x=x.reshape(50,1)label_y=label_y.reshape(50,1)train_data...

生成数据import numpy as npimport matplotlib.pyplot as pltc0=np.random.normal(6,1,100)c1=np.random.normal(3,1,100)c2=np.random.normal(3,1,100)c3=np.random.normal(6,1,100)c0y=np.zeros(100)c1y=np....

from sklearn.datasets import make_blobsfrom matplotlib import pyplotdata,label = make_blobs(n_samples=1000,n_features=2,centers=5)# 绘制样本显示pyplot.scatter(data[:,0],data[:,1],c=label)pyplot.sho...

from sklearn.datasets.samples_generator import make_classificationX, y = make_classification(n_samples=200, n_features=4, n_informative=2, n_redundant=2, n_classes=2, n_clusters_per_class=2, sc...

from sklearn.cluster import KMeanskmeans = KMeans(n_clusters = 3)kmeans.fit(x)y_kmeans = kmeans.predict(x)plt.scatter(x[:,0], x[:,1], c = y_kmeans, cmap=‘Dark2’, s=50, alpha=0.5, marker=‘x’)centr...

1 获得数据from sklearn import datasetsiris=datasets.load_iris()iris_x=iris.datairis_y=iris.target2 处理数据，数据分割成测试集和训练集from sklearn.model_selection import train_test_splitX_train,X_test,y_train,y...