<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[GinPd的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/GinPd</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; GinPd]]></copyright><item><title><![CDATA[【论文阅读】2023_DifFSS: Diffusion Model for Few-Shot Semantic Segmentation （arxiv）]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/132650053</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/132650053</guid><author>GinPd</author><pubDate>Sun, 03 Sep 2023 14:32:33 +0800</pubDate><description><![CDATA[（如autoregressive models、GAN）DDPM: Denoising Diffusion Probabilistic Models中的ControlNet。是通过diffusion扩散模型生成图像将支持集样本从1shot增加到n+1shot，也就是生成辅助support图像。用于小样本语义分割的扩散模型。]]></description><category></category></item><item><title><![CDATA[【解决】报错：部署Deformable-DETR环境问题]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/130019811</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/130019811</guid><author>GinPd</author><pubDate>Fri, 07 Apr 2023 20:25:42 +0800</pubDate><description><![CDATA[阅读readme发现需要在 model/opts/下 sh make.sh 安装环境。最初始时报错：没有 MSDeformAttnFunction 这个module。又报了一大长段错，找到一个错误行。在配置文件中添加如下一行。保存后source更新。再运行程序，问题解决。]]></description><category></category></item><item><title><![CDATA[【论文阅读】APANet：用于小样本分割的自适应原型对齐网络]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/129877516</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/129877516</guid><author>GinPd</author><pubDate>Fri, 31 Mar 2023 13:02:33 +0800</pubDate><description><![CDATA[本文给出了对图像背景的一种处理方法，即将背景视作要分割的部分（1-Ms），利用可知类和不可知类做特征对齐，类似于BAM，去除不想分割的部分。我们已经知道支持和查询前景相似，但背景可能完全不同，所以文中直接对查询背景分析，使用K-means聚类。其中特征对齐部分，背景分割分支学习参数与前景分支共享，提供了一种利用负样本的方法。]]></description><category></category></item><item><title><![CDATA[【小样本分割论文】顶会链接及FSS论文]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/129877340</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/129877340</guid><author>GinPd</author><pubDate>Fri, 31 Mar 2023 12:48:10 +0800</pubDate><description><![CDATA[顶会论文链接即小样本分割论文]]></description><category></category></item><item><title><![CDATA[【pytorch报错】tensorboard无法显示问题]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/129177221</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/129177221</guid><author>GinPd</author><pubDate>Thu, 23 Feb 2023 10:55:22 +0800</pubDate><description><![CDATA[出现tensorboard无法可视化log问题。试了网上一些提供方案，还是没能解决。]]></description><category></category></item><item><title><![CDATA[【论文阅读】Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/127246474</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/127246474</guid><author>GinPd</author><pubDate>Mon, 10 Oct 2022 16:17:55 +0800</pubDate><description><![CDATA[Freeze 预训练的backbone（(冻结主干网络的参数)已经成为FSS任务中预防 over-fitting 的标准范式，但本文提出一种新范式，即fine-tuning a small part of parameters in the backbone。]]></description><category></category></item><item><title><![CDATA[【论文阅读】MaskSplit:用于小样本语义分割的自监督元学习]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/126643862</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/126643862</guid><author>GinPd</author><pubDate>Thu, 01 Sep 2022 15:14:41 +0800</pubDate><description><![CDATA[核心思想一是缺少样本就通过图像增强创造图像，二是在增强的基础上使用一分为二的掩膜分割方法构造支持和查询对。]]></description><category></category></item><item><title><![CDATA[【计算机网络】第五章 网络层]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123514534</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123514534</guid><author>GinPd</author><pubDate>Tue, 15 Mar 2022 22:38:53 +0800</pubDate><description><![CDATA[]]></description><category></category></item><item><title><![CDATA[【计算机网络】第四章 数据链路层]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123514491</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123514491</guid><author>GinPd</author><pubDate>Tue, 15 Mar 2022 22:36:37 +0800</pubDate><description><![CDATA[]]></description><category></category></item><item><title><![CDATA[【计算机网络】第三章 数据通信基础与物理层]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123514458</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123514458</guid><author>GinPd</author><pubDate>Tue, 15 Mar 2022 22:34:54 +0800</pubDate><description><![CDATA[]]></description><category></category></item><item><title><![CDATA[【计算机网络】第二章 网络协议和网络体系结构]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123514364</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123514364</guid><author>GinPd</author><pubDate>Tue, 15 Mar 2022 22:30:53 +0800</pubDate><description><![CDATA[]]></description><category></category></item><item><title><![CDATA[【计算机网络】第一章 概论]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123514064</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123514064</guid><author>GinPd</author><pubDate>Tue, 15 Mar 2022 22:27:59 +0800</pubDate><description><![CDATA[]]></description><category></category></item><item><title><![CDATA[Python 猜数字小游戏]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123513561</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123513561</guid><author>GinPd</author><pubDate>Tue, 15 Mar 2022 21:57:11 +0800</pubDate><description><![CDATA[import random #生成随机数
secret=random.randint(1,100)
print("猜数字游戏！1—100的整数，你最多可以猜9次")
tries=1
while tries&lt;=9:  
  guess=int(input("1-100的整数,第%d次猜,请输入:"%(tries)))
  if guess==secret:
    print("答对，本轮游戏结束\n")
    break
  elif guess&gt;secret:
    print("大了")]]></description><category></category></item><item><title><![CDATA[HDLBits-verilog刷题]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123388331</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123388331</guid><author>GinPd</author><pubDate>Wed, 09 Mar 2022 21:42:15 +0800</pubDate><description><![CDATA[Vectors
Four-input gates
module top_module(  
input [3:0] in,  
output out_and,  
output out_or,  
output out_xor  
);  
assign out_and = &amp;in;  
assign out_or = |in;  
assign out_xor = ^in;  
  
endmodule  

Vector concatenation operator
module top_mod]]></description><category></category></item><item><title><![CDATA[HDLBits-verilog刷题]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/123194320</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/123194320</guid><author>GinPd</author><pubDate>Mon, 28 Feb 2022 22:33:41 +0800</pubDate><description><![CDATA[Vectors
Gates4
module top_module(  
input [3:0] in,  
output out_and,  
output out_or,  
output out_xor  
);  
assign out_and = &amp;in;  
assign out_or = |in;  
assign out_xor = ^in;  
  
endmodule  

Vector3
module top_module (
    input [4:0] a, b, c, d]]></description><category></category></item><item><title><![CDATA[python 实验1.3]]></title><link>https://blog.youkuaiyun.com/GinPd/article/details/112300274</link><guid>https://blog.youkuaiyun.com/GinPd/article/details/112300274</guid><author>GinPd</author><pubDate>Wed, 06 Jan 2021 23:48:42 +0800</pubDate><description><![CDATA[python 实验1.3
题目：编写函数：随机产生包含 n 个整数的列表，返回一个元组，其中第一个 元素为所有参数的平均值，其他元素为所有参数中大于平均值的整数。 例如： 随机产生长度为 3 的列表[2, 1, 3]，输出为（2.0, 3）
import random
def list_avg(n):
    a_list =  random.sample(range(51),n)
    avg = sum(a_list)/n
    l = [avg]
    for i in a_list:
    ]]></description><category></category></item></channel></rss>