- 博客(21)
- 资源 (1)
- 收藏
- 关注
RSS订阅原创 leetcode 每日一题 单词拆分
首先定义一个长度为len(s)+1的全为False列表，然后将索引为0的元素设为True。然后遍历列表的索引，也就是从1到len（s）+1遍历worddict，如果列表的索引>=word的长度，那么让此索引位置的元素不变或者当word存在与字符串中时s[j - len(word):j])等于dp[j - len(word)]，由于第一个元素为true，所以如果if后的假设以及word == s[j - len(word):j]都正确的话，那么true会传递下去，直到最后一位，所以最后的dp[.
2022-02-23 23:37:13
197
原创 python把文本标签转换为数字标签
def label_num_list(train_path): train=pd.read_csv(train_path) n_classes = len(train) leaves_labels = sorted(list(set(train['label']))) class_to_num = dict(zip(leaves_labels, range(n_classes))) train_label=train['label'] num_train_l.
2022-02-07 10:22:57
3348
原创 pytorch本地图片数据集加载成字典
from torchvision.transforms import transformsimport pandas as pdfile_path=os.path.join(r'E:\python存储\leaves\images')Image_list=[]labels_list=[]a={}for i in range(0,18352): image_path=os.path.join(r'E:\python存储\leaves\images','{}.jpg'.format(i).
2022-02-06 20:55:30
1493
原创 AlexNet pytorch代码实现
import torchfrom torch import nnfrom d2l import torch as d2lnet=nn.Sequential( nn.Conv2d(1,96,kernel_size=11,stride=4,padding=1),nn.ReLU(), nn.MaxPool2d(kernel_size=3,stride=2), nn.Conv2d(96,128*2,kernel_size=5,padding=2),nn.ReLU(),.
2022-02-01 18:52:13
1459
原创 pytorch检查模型各部分输入与输出
import torchfrom torch import nnfrom d2l import torch as d2lclass Reshape(torch.nn.Module): def forward(selfself,x): return x.view(-1,1,28,28)net=torch.nn.Sequential(Reshape(), nn.Conv2d(1,6,kernel_size=5,padding=.
2022-01-30 21:01:20
3369
1
原创 torch池化层基本原理代码
# 池化层的正向传播import torchfrom torch import nnfrom d2l import torch as d2ldef pool2d(X,pool_size,mode='max'): p_h,p_w=pool_size Y=torch.zeros((X.shape[0]-p_h+1,X.sjape[1]-p_w+1)) for i in range(Y.shape[0]): for j in range(Y.shape[1]):.
2022-01-29 23:13:26
703
1
原创 pytorch卷积层代码原理实现
import torchfrom torch import nnfrom d2l import torch as d2ldef corr2d(X,K): # 计算二维互相关运算 h,w=K.shape Y=torch.zeros(X.shape[0]-h+1,X.shape[1]-w+1) for i in range(Y.shape[0]): for j in range(Y.shape[1]): Y[i,j]=(X[i:i.
2022-01-28 16:15:15
1260
原创 python爬取携程评论
import requestsimport jsonimport timepagesize=300headers = { 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/94.0.4606.71 Safari/537.36',}posturl = "https://m.ctrip.com/restapi/soa2/13444/j.
2022-01-26 01:24:12
1724
8
原创 pytorch参数管理
import torchfrom torch import nnnet=nn.Sequential(nn.Linear(4,8),nn.ReLU(),nn.Linear(8,1))x=torch.rand(size=(2,4))y=net(x)print(y)print(net[2].state_dict())print(type(net[2].bias))print(net[2].bias)print(net[2].bias.data)print(net[2].weight.
2022-01-24 20:15:16
913
1
原创 pytorch神经网络基础
import torchfrom torch import nnfrom torch .nn import functional as F# 定义了一个特殊的module# net = nn.Sequential(nn.Linear(20,256),nn.ReLU(),nn.Linear(256,10))x=torch.rand(2,20)# print(net(x))# 感知机class MLP(nn.Module): #继承module父类，module有两个非常重要的函数 .
2022-01-18 21:33:48
971
原创 dropout代码实现原理
def dropout_layer(x,dropout): assert 0 <= dropout<=1 if dropout==1: return torch.zeros_like(x) if dropout==0: return x mask=(torch.randn(x.shape)>dropout).float() return mask*x /(1.0-dropout)
2022-01-16 22:23:07
746
1
原创 多层感知机fashion mnist
import torchfrom matplotlib import pyplot as pltfrom torch import nnfrom d2l import torch as d2lnet=nn.Sequential(nn.Flatten(),nn.Linear(784,256),nn.ReLU(), nn.Linear(256,10))def init_weights(m): if type(m)==nn.Linear: .
2022-01-07 22:58:10
775
原创 softmax代码实现，代码原理供参考
import torchfrom IPython import displayfrom d2l import torch as d2limport torch.utils.data.dataloaderfrom matplotlib import pyplot as plt# import pylabbatch_size=64train_iter,test_iter=d2l.load_data_fashion_mnist(batch_size) #训练集和测试集的迭代器num_inpu.
2022-01-06 23:23:21
1029
原创 线性回归与梯度下降的代码实现
import randomimport torchfrom d2l import torch as d2limport matplotlib.pyplot as plt# 构造数据集def syn_data(w,b,num_examples): x=torch.normal(0,1,(num_examples,len(w)))#均值为0，方差为一的随机数 y=torch.matmul(x,w)+b y+=torch.normal(0,0.01,y.shape) .
2022-01-04 22:27:43
404
原创 knn代码实现
from sklearn.neighbors import KNeighborsClassifierimport pandas as pdimport matplotlib.pyplot as pltimport seaborn as snsimport numpy as npvalues=np.array([ [0.2,0.2], [0.3,0.3], [0.4,0.4], [0.8,0.8]])labels=['A','B','C','D']# df=.
2022-01-03 22:51:29
482
原创 pytorch-隐式自动求导
import torchx=torch.arange(4.0)print(x)x.requires_grad_(True)print(x.grad)y=2*torch.dot(x,x)print(y)y.backward()print(x.grad)print(x.grad==4*x)# 默认情况下，pytorch会累计梯度，我们需要清除之前的值x.grad.zero_()y=x.sum()y.backward()print(x.grad)x.grad.zero_.
2022-01-03 22:34:09
602
原创 卷积神经网络测试代码-cifar10
import torchimport torchvisionfrom PIL import Imagefrom torch import nnfrom pytorch.小土堆.moudle import liimage_path = '../g.jpg'image = Image.open(image_path)print(image)transform = torchvision.transforms.Compose([torchvision.transforms.Resize((.
2022-01-03 22:23:19
1180
原创 卷积神经网络-训练代码（cifar10）
import torch.optim.sgdimport torchvision# 准备数据集from torch import nnfrom torch.utils.data import DataLoaderfrom torch.utils.tensorboard import SummaryWriter# 定义训练的设备device=torch.device('cpu')writer=SummaryWriter('./log_train')train_data=torchv.
2022-01-03 22:22:06
830
2
原创 Bert自学笔记1
传统Rnn有什么问题：每一个下一步都会用到前面的结果，不独立，不能并行计算，消耗算力大，所以层数较少，不能像卷积那样有很多层。Transformer可以作并行。Rnn中的注意力机制，即给每个词分配重要性权重，让网络知道哪个重要，哪个不重要。此即是ATTention。传统的word2vec用词向量表示文本特征，救活了nlp。当训练传统word2vec时，先训练好一个词向量，再对语料库中的文本转换为词向量，传入rnn，输入到词向量的词不会变，但不同的词，在不同语义中可能不一样。同样一个词在不同句
2021-12-24 11:04:08
491
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人