- 博客(14)
- 资源 (3)
- 收藏
- 关注
原创 基于pytorch的FCN网络简单实现
参考知乎专栏实现FCN网络https://zhuanlan.zhihu.com/p/32506912import torchfrom torch import nnimport torch.nn.functional as fimport torchvisionimport torchvision.transforms as tfsfrom torch.utils.data import DataLoaderfrom torch.autograd import Variableimport
2020-07-29 11:48:48 10117 38
原创 使用pytorch进行迁移学习模型下载失败解决办法
使用pytorch进行迁移学习的时候,我们需要下载预训练的模型,但是这个模型通常很大,如果在代码中在线下载的话,很可能会中断,并且一中断之前也就白下载了,这篇文章里我介绍一种离线使用预训练模型的方法。所谓离线使用预训练模型的方法,实际上就是使用浏览器将模型下载下来(通常浏览器下载会比较稳定,并且如果下载中断还能恢复),下面给出各种模型的下载地址,只需要将对应的url键入到浏览器中就可以建立下载1. Resnet: model_urls = { 'resnet18': 'https:
2020-07-26 20:25:12 6182
原创 神经网络中转置卷积上采样与反最大池化上采样的对比
目前,CNN卷积神经网络中,图像上采样通常有两种方法,分别是FCN网络使用的转置卷积和segnet中使用的反最大池化。我们首先来看转置卷积的上采样方法,关于转置卷积的原理,可以看下面这篇文章https://blog.csdn.net/lanadeus/article/details/82534425我们知道,卷积网络输入输出尺寸的关系如下nout=nin−kernel+2∗paddingstride+1 n_{out} = \frac{n_{in}-kernel+2*padding} {stride
2020-07-23 20:51:55 822 2
原创 1102 Invert a Binary Tree (25分)
The following is from Max Howell @twitter:Google: 90% of our engineers use the software you wrote (Homebrew), but you can't invert a binary tree on a whiteboard so fuck off.Now it's your turn to prove that YOU CAN invert a binary tree!Input Specifi
2020-07-16 19:16:10 187
原创 pytorch CNN CIFAR10数据集识别
尝试使用深层结构进行CIFAR10的识别import torchimport torchvisionimport torchvision.transforms as transformsBATCH_SIZE = 64EPOCHES = 50NUM_WORKERS = 4LEARNING_RATE = 0.005# 数据转换transform = transforms.Compose( [transforms.ToTensor(), transforms.Normali
2020-07-16 14:55:24 2032
原创 1021 Deepest Root (25分)
A graph which is connected and acyclic can be considered a tree. The height of the tree depends on the selected root. Now you are supposed to find the root that results in a highest tree. Such a root is calledthe deepest root.Input Specification:Each .
2020-07-15 15:52:32 242
原创 卷积动画详解
卷积算法这篇文章是深度学习中卷积算法的动画演示本教程的代码和图像可按以下规定免费使用许可并受适当的署名:[1] Vincent Dumoulin, Francesco Visin - A guide to convolution arithmeticfor deep learning(BibTeX)卷积动画蓝色的图像是输入图像,青色的图像是输出图像没有边界填充,无跨步(No padding, no strides)任意的边界填充,无跨步()半填充,无跨步(Half pa
2020-07-14 10:43:31 4102
原创 1020 Tree Traversals (25分)
Suppose that all the keys in a binary tree are distinct positive integers. Given the postorder and inorder traversal sequences, you are supposed to output the level order traversal sequence of the corresponding binary tree.Input Specification:Each inpu
2020-07-13 16:20:13 361
原创 pytorch全连接神经网络进行MNIST识别
这个例程使用全连接神经网络进行MNIST识别import numpy as npimport torchfrom torchvision.datasets import mnistfrom torch import nnfrom torch.autograd import Variabledef data_tf(x): x = np.array(x, dtype="float32")/255 x = (x-0.5)/0.5 x = x.reshape((-1))
2020-07-13 11:42:22 1657 1
原创 1103 Integer Factorization (30分)
TheK−Pfactorization of a positive integerNis to writeNas the sum of theP-th power ofKpositive integers. You are supposed to write a program to find theK−Pfactorization ofNfor any positive integersN,KandP.Input Specification:Each input ...
2020-07-10 10:14:15 308
原创 1037 Magic Coupon
The magic shop in Mars is offering some magic coupons. Each coupon has an integerNprinted on it, meaning that when you use this coupon with a product, you may getNtimes the value of that product back! What is more, the shop also offers some bonus produ...
2020-07-10 08:33:48 153
原创 1033 To Fill or Not to Fill
1033To Fill or Not to Fill(25分)With highways available, driving a car from Hangzhou to any other city is easy. But since the tank capacity of a car is limited, we have to find gas stations on the way from time to time. Different gas station may give di..
2020-07-09 17:07:38 269
原创 1003 Emergency (25分)
As an emergency rescue team leader of a city, you are given a special map of your country. The map shows several scattered cities connected by some roads. Amount of rescue teams in each city and the length of each road between any pair of cities are marked
2020-07-08 20:39:47 176
原创 python构建决策树
决策树是一常见的机器学习算法,本例程将参考《机器学习实践》中的代码完成决策树算法中,信息熵被用来定义数据的纯度。假定当前样本集合D中第k类样本所占比例为$ p_k $,则有 Ent(D)=−∑k=1npklog2pk Ent(D)=-\sum_{k=1}^n p_k \log_2 p_k Ent(D)=−k=1∑npklog2pk我们认为Ent(D)的值越小,则D的纯度越高。假定离散属性a有V个可能的取值$ { a^1, a^2, …, a^V } $ ,若用a对样本集D进行划分,则会产
2020-07-03 12:22:06 2148
c++背词宝应用 exe文件
2018-09-05
c++背词宝应用
2018-09-05
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人