<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[xiyou__的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/xiyou__</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; xiyou__]]></copyright><item><title><![CDATA[tkinter实现一个GUI界面-快速入手]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/137679921</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/137679921</guid><author>xiyou__</author><pubDate>Wed, 19 Jun 2024 10:37:14 +0800</pubDate><description><![CDATA[适用于新手需要为自己的程序实现一个前端界面。]]></description><category></category></item><item><title><![CDATA[Pandas 常用操作（持续更新）]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/132862063</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/132862063</guid><author>xiyou__</author><pubDate>Mon, 17 Jun 2024 14:38:10 +0800</pubDate><description><![CDATA[DataFrame在实际使用中常用的操作代码。]]></description><category></category></item><item><title><![CDATA[jupyter快速实现单标签及多标签多分类的文本分类BERT模型]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/134603753</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/134603753</guid><author>xiyou__</author><pubDate>Fri, 05 Jan 2024 16:02:59 +0800</pubDate><description><![CDATA[jupyter快速实现单标签及多标签多分类的文本分类BERT模型。]]></description><category></category></item><item><title><![CDATA[SentiBank Dector上手指南]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/130215200</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/130215200</guid><author>xiyou__</author><pubDate>Sat, 29 Apr 2023 15:00:27 +0800</pubDate><description><![CDATA[​官网链接：https://www.ee.columbia.edu/ln/dvmm/vso/download/sentibank.htmlSentiBank Detector可以抽取图片中的形容词-名词对，之前一直看到，这次复现模型才第一次用到，上手的时候有点手足无措，因为官网在如何使用方面说的不是很清楚，但是用完发现也挺简单，这里记录一下。]]></description><category></category></item><item><title><![CDATA[latex入门指南：插入图片、表格、公式方法一览]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/129163491</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/129163491</guid><author>xiyou__</author><pubDate>Thu, 09 Mar 2023 11:41:31 +0800</pubDate><description><![CDATA[latex论文编辑指南]]></description><category></category></item><item><title><![CDATA[GPT系列学习笔记：GPT、GPT2、GPT3]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/126388844</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/126388844</guid><author>xiyou__</author><pubDate>Fri, 19 Aug 2022 11:22:58 +0800</pubDate><description><![CDATA[GPT、GPT2、GPT3的区别和联系。]]></description><category></category></item><item><title><![CDATA[中文词向量word2vec计算相似度]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/125634411</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/125634411</guid><author>xiyou__</author><pubDate>Wed, 06 Jul 2022 15:59:26 +0800</pubDate><description><![CDATA[基于word2vec词向量计算中文词语之间的相似度。]]></description><category></category></item><item><title><![CDATA[TensorboardX：如何在pytorch上画loss曲线图 | TensorboardX入门使用指南]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/125618290</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/125618290</guid><author>xiyou__</author><pubDate>Tue, 05 Jul 2022 14:53:57 +0800</pubDate><description><![CDATA[TensorboardX入门使用指南]]></description><category></category></item><item><title><![CDATA[DL论文词汇积累（持更）]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/124976480</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/124976480</guid><author>xiyou__</author><pubDate>Sat, 25 Jun 2022 14:24:11 +0800</pubDate><description><![CDATA[深度学习领域论文词汇的阅读与积累。]]></description><category></category></item><item><title><![CDATA[WordPiece 和 BPE 的区别]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/124173953</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/124173953</guid><author>xiyou__</author><pubDate>Tue, 21 Jun 2022 17:29:30 +0800</pubDate><description><![CDATA[区分WordPiece和BPE


]]></description><category></category></item><item><title><![CDATA[百面机器学习阅读笔记（二）：集成学习]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/124333381</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/124333381</guid><author>xiyou__</author><pubDate>Wed, 27 Apr 2022 23:02:34 +0800</pubDate><description><![CDATA[百面机器学习第12章：集成学习 部分复习内容。]]></description><category></category></item><item><title><![CDATA[百面机器学习阅读笔记（一）：模型评估]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/123776834</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/123776834</guid><author>xiyou__</author><pubDate>Wed, 30 Mar 2022 11:42:43 +0800</pubDate><description><![CDATA[阅读《百面机器学习》第二章部分内容。
]]></description><category></category></item><item><title><![CDATA[论文阅读：An Empirical Study of Training End-to-End Vision-and-Language Transformers]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/123134897</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/123134897</guid><author>xiyou__</author><pubDate>Fri, 25 Feb 2022 18:52:40 +0800</pubDate><description><![CDATA[摘要
提出Multimodal End-to-end TransformER framework，即METER，通过这个框架研究如何以一个端到端的方式( in an end-to-end manner )设计和预训练一个完全基于transformer的视觉语言模型。

端到端指的是输入是原始数据，输出是最后的结果，整个模型过程相当于黑箱操作；相反，非端到端的方法，数据的预处理部分是单独的模块，需要人工处理（如人工标注数据，人工提取图片特征）（笔者注）

具体地，模型从多个维度被解剖(dissect)为：
视]]></description><category></category></item><item><title><![CDATA[Jupyter Notebook对数据集进行数据分析 数据统计(含：样本数量/样本长度/样本长度分布图/示例样本打印)]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/123130346</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/123130346</guid><author>xiyou__</author><pubDate>Fri, 25 Feb 2022 12:30:15 +0800</pubDate><description><![CDATA[以下内容可直接以写入.ipynb文件的形式，放入服务器上任意已准备好数据集的文件夹下：

导入模块

import os
import json
import re
from collections import defaultdict


绘图函数

def draw(dic): #输入样本数量统计字典
    import numpy as np
    import matplotlib.pyplot as plt
    from matplotlib.ticker import MaxNLocat]]></description><category></category></item><item><title><![CDATA[bert的变形：RoBERTa 和 BERT 的异同]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/122192900</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/122192900</guid><author>xiyou__</author><pubDate>Tue, 28 Dec 2021 16:10:49 +0800</pubDate><description><![CDATA[RoBERTa和BERT对比。]]></description><category></category></item><item><title><![CDATA[谷歌翻译 google translation API github开源 实践]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/122171005</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/122171005</guid><author>xiyou__</author><pubDate>Mon, 27 Dec 2021 14:46:24 +0800</pubDate><description><![CDATA[google translation API 无限量翻译]]></description><category></category></item><item><title><![CDATA[多模态bert的变形：VisualBERT 和 BERT 的异同]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/122005851</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/122005851</guid><author>xiyou__</author><pubDate>Sat, 18 Dec 2021 16:29:53 +0800</pubDate><description><![CDATA[BERT
VisualBERT
备注




全称
Bidirectional Encoder Representations from Transformers




区别于传统Transformer
使用 双向 (bidirectional) self-attention
集成了 BERT, Transformer-based model for NLP, object proposals systems such as Faster-RCNN



预训练数据集
wiki
COCO ima...]]></description><category></category></item><item><title><![CDATA[跑通mmf：visualBert实现过程]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/121792839</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/121792839</guid><author>xiyou__</author><pubDate>Mon, 13 Dec 2021 10:14:23 +0800</pubDate><description><![CDATA[使用代码github链接：https://github.com/di-dimitrov/propaganda-techniques-in-memes
该代码相当在mmf-master上展开自己的任务，之前想直接跑通facebook的MMF，但是在环境配置问题上屡屡碰壁，后来被推荐了这套代码，才跑通了visualBert模型。
实现过程
1 将项目下载并存放到服务器
2 Install MMF


Prerequisites - generating image caption features for Vi]]></description><category></category></item><item><title><![CDATA[使用os.path.join无效/不起作用的原因]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/121627712</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/121627712</guid><author>xiyou__</author><pubDate>Tue, 30 Nov 2021 10:57:28 +0800</pubDate><description><![CDATA[os.path.join(a,b)在以下情况下会起不到连接作用，并返回b:
a,b中存在一个为绝对路径，即不是patha/pathb的形式，而是/patha/pathb的形式

]]></description><category></category></item><item><title><![CDATA[解决合并压缩包分卷无法解压 错误信息：文件格式未知或者压缩文件数据已经损坏]]></title><link>https://blog.youkuaiyun.com/xiyou__/article/details/121531258</link><guid>https://blog.youkuaiyun.com/xiyou__/article/details/121531258</guid><author>xiyou__</author><pubDate>Thu, 25 Nov 2021 10:12:41 +0800</pubDate><description><![CDATA[​
问题起因：
我想要获取Semantic Image-Text-Classes - Datasets - Forschungsdaten-Repositorium der LUH (uni-hannover.de)中的数据集，该数据集的train被分为了49个分卷，当我按照他的要求下载了所有的分卷并执行cat train.tar.part* &gt; train_concat.tar合并分卷后，却无法打开或解压压缩包，报错内容如下（打开压缩包时弹出错误）：

（文件格式未知或者压缩文件数据已损坏。）
最后]]></description><category></category></item></channel></rss>