<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[真理无穷，进一步有进一步的欢喜]]></title><description><![CDATA[]]></description><link>https://blog.csdn.net/qq_41683065</link><language>zh-cn</language><generator>https://blog.csdn.net/</generator><copyright><![CDATA[Copyright &copy; qq_41683065]]></copyright><item><title><![CDATA[[Ubuntu] download and install ChromeDrive]]></title><link>https://blog.csdn.net/qq_41683065/article/details/134105586</link><guid>https://blog.csdn.net/qq_41683065/article/details/134105586</guid><author>qq_41683065</author><pubDate>Sun, 29 Oct 2023 20:14:14 +0800</pubDate><description><![CDATA[【代码】[Ubuntu] download and install ChromeDrive。]]></description><category></category></item><item><title><![CDATA[[Ubuntu] set screen resolution]]></title><link>https://blog.csdn.net/qq_41683065/article/details/127115425</link><guid>https://blog.csdn.net/qq_41683065/article/details/127115425</guid><author>qq_41683065</author><pubDate>Thu, 29 Sep 2022 22:52:37 +0800</pubDate><description><![CDATA[rt]]></description><category></category></item><item><title><![CDATA[[Python / PyTorch] debug backward()]]></title><link>https://blog.csdn.net/qq_41683065/article/details/123551427</link><guid>https://blog.csdn.net/qq_41683065/article/details/123551427</guid><author>qq_41683065</author><pubDate>Thu, 17 Mar 2022 15:22:34 +0800</pubDate><description><![CDATA[问题描述
在自定义Loss的中，其backward()函数不支持在PyCharm中进行断点调试
因此需要以其他方式进行断点调试
解决方案
参考：Is there a way to debug the backward method of Function class?
使用pdb，在对应位置增加如下代码：
        import pdb
        pdb.set_trace()

具体pdb命令，参考：10分钟教程掌握Python调试器pdb

...]]></description><category></category></item><item><title><![CDATA[[PyTorch] 可视化tensor图像]]></title><link>https://blog.csdn.net/qq_41683065/article/details/122649756</link><guid>https://blog.csdn.net/qq_41683065/article/details/122649756</guid><author>qq_41683065</author><pubDate>Sun, 23 Jan 2022 15:17:28 +0800</pubDate><description><![CDATA[PyTorch提供了直接对tensor可视化的接口函数：
plt.imshow(transforms.ToPILImage()(image), interpolation="bicubic")
transforms.ToPILImage()(image).show() # Alternatively

参考：How do I display a single image in PyTorch?

]]></description><category></category></item><item><title><![CDATA[[PyTorch] reverse the normalization]]></title><link>https://blog.csdn.net/qq_41683065/article/details/122649316</link><guid>https://blog.csdn.net/qq_41683065/article/details/122649316</guid><author>qq_41683065</author><pubDate>Sun, 23 Jan 2022 11:55:35 +0800</pubDate><description><![CDATA[有时候在程序中需要对图像做可视化，但是读取进来的tensor已经经过了normalization，为小数，甚至可能出现负值
为了可视化图像，必须对将tensor通过逆标准化获得原始的tensor
具体方法参考 [Feature Request] Un-Normalize Image Tensor
mean = torch.tensor([1, 2, 3], dtype=torch.float32)
std = torch.tensor([2, 2, 2], dtype=torch.float32)

nor]]></description><category></category></item><item><title><![CDATA[[debug] PyTorch报错：ConnectionResetError: [Errno 104] Connection reset by peer]]></title><link>https://blog.csdn.net/qq_41683065/article/details/122643637</link><guid>https://blog.csdn.net/qq_41683065/article/details/122643637</guid><author>qq_41683065</author><pubDate>Sat, 22 Jan 2022 21:07:00 +0800</pubDate><description><![CDATA[问题描述：
使用PyTorch 1.10.0，训练报错：
ConnectionResetError: [Errno 104] Connection reset by peer

问题解析
参见pytorch的issue

I believe the issue is only triggered for the case that both
persistent_workers and pin_memory are turned on and iteration is
terminated at the t]]></description><category></category></item><item><title><![CDATA[[debug] Expected to have finished reduction in the prior iteration before starting a new one.]]></title><link>https://blog.csdn.net/qq_41683065/article/details/122442379</link><guid>https://blog.csdn.net/qq_41683065/article/details/122442379</guid><author>qq_41683065</author><pubDate>Tue, 11 Jan 2022 21:48:36 +0800</pubDate><description><![CDATA[问题描述
训练网络时出现错误：
RuntimeError: Expected to have finished reduction in the prior iteration before starting a new one.

原因分析
参考该issue

I met the same issue.
But i solved it.
The reason is that in my model class, I define a fpn module with 5 level output featu]]></description><category></category></item><item><title><![CDATA[[debug] RuntimeError: “nll_loss_forward_reduce_cuda_kernel_2d_index“ not implemented for ‘Int‘]]></title><link>https://blog.csdn.net/qq_41683065/article/details/122441927</link><guid>https://blog.csdn.net/qq_41683065/article/details/122441927</guid><author>qq_41683065</author><pubDate>Tue, 11 Jan 2022 21:20:56 +0800</pubDate><description><![CDATA[问题描述
在使用nn.CrossEntropyLoss时出现报错：
loss_id = self.IDLoss(id_preds, id_targets)

Traceback (most recent call last):
  File "/home/hgx/pycharm-2020.1.3/plugins/python/helpers/pydev/_pydevd_bundle/pydevd_exec2.py", line 3, in Exec
    exec(exp, global_vars, lo]]></description><category></category></item><item><title><![CDATA[python 取列表（数组）偶数和奇数位置的值]]></title><link>https://blog.csdn.net/qq_41683065/article/details/118877680</link><guid>https://blog.csdn.net/qq_41683065/article/details/118877680</guid><author>qq_41683065</author><pubDate>Sun, 18 Jul 2021 15:14:51 +0800</pubDate><description><![CDATA[python中的列表[]等于其他语言中的数组[]
取列表索引取值，也就是列表的切片：
list[i:j] 就是从列表的索引 i 到索引j 个的值；列表的索引第一位是从0开始的
list[i:j:2] 一样取i 到 j 但加入了步长 这里步长为2；也就是取每次索引位置开始+2的值
list[::2 ] 就是取奇数位 这里的 i j 我们省略的话就是默认数组最开头到结尾
list[1::2] 这里缺省了j 但是i定义了1 也就是从数组第二个数开始取 ，所以这个是取偶数位


...]]></description><category></category></item><item><title><![CDATA[[debug] 延长Linux中sudo密码在终端的有效时间]]></title><link>https://blog.csdn.net/qq_41683065/article/details/118827775</link><guid>https://blog.csdn.net/qq_41683065/article/details/118827775</guid><author>qq_41683065</author><pubDate>Fri, 16 Jul 2021 22:23:03 +0800</pubDate><description><![CDATA[参考：如何在 Linux 中让 sudo 密码会话的超时更长些

问题描述
在第一次执行 sudo 命令后，默认情况下密码将保持 15 分钟。
某些需要长时间执行且需要sudo权限的命令，如果超过15min的时间上线就会报错。
所以需要延长sudo密码在终端的有效时间。
解决方法
首先使用 sudo  命令以超级用户权限打开 /etc/sudoers 文件，如下所示：
sudo gedit /etc/sudoers

接着添加下面的默认值，这意味着 sudo 密码提示将会在用户使用 sudo 20 分钟后.]]></description><category></category></item><item><title><![CDATA[[debug] pycharm在debug模式下一直显示collecting data]]></title><link>https://blog.csdn.net/qq_41683065/article/details/118704707</link><guid>https://blog.csdn.net/qq_41683065/article/details/118704707</guid><author>qq_41683065</author><pubDate>Tue, 13 Jul 2021 16:53:36 +0800</pubDate><description><![CDATA[在setting --&gt; Build, Execution, Deployment --&gt; Python Debugger下，选中第三项"Gevent compatible"即可


]]></description><category></category></item><item><title><![CDATA[[Linux + 深度学习]Ubuntu18.04 深度学习环境配置]]></title><link>https://blog.csdn.net/qq_41683065/article/details/118640329</link><guid>https://blog.csdn.net/qq_41683065/article/details/118640329</guid><author>qq_41683065</author><pubDate>Sat, 10 Jul 2021 20:48:31 +0800</pubDate><description><![CDATA[contact author: Han GuangXin

NVIDIA显卡驱动
安装NVIDIA显卡驱动
software&amp;update ==&gt; 附加驱动 ==&gt; 选择对应驱动 ==&gt; 应用更改
ubuntu-drivers devices



禁用Ubuntu自带的nouveau驱动
桌面终端输入：
sudo gedit /etc/modprobe.d/blacklist-nouveau.conf

打开后为空白文件，加入下面两行并保存：
blacklist nouveau.]]></description><category></category></item><item><title><![CDATA[[debug] RuntimeError: CUDA error: no kernel image is available for execution on the device]]></title><link>https://blog.csdn.net/qq_41683065/article/details/118575107</link><guid>https://blog.csdn.net/qq_41683065/article/details/118575107</guid><author>qq_41683065</author><pubDate>Thu, 08 Jul 2021 16:05:01 +0800</pubDate><description><![CDATA[问题描述
运行程序时出现报错：
RuntimeError: CUDA error: no kernel image is available for execution on the device (launch_kernel at /opt/conda/conda-bld/pytorch_1556653114079/work/aten/src/ATen/native/cuda/Loops.cuh:72)

问题原因
当前使用的硬件算力不支持相应的pytorch函数，需要更换pytorch到一个合适的版本。]]></description><category></category></item><item><title><![CDATA[[debug] you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake ……]]></title><link>https://blog.csdn.net/qq_41683065/article/details/110200453</link><guid>https://blog.csdn.net/qq_41683065/article/details/110200453</guid><author>qq_41683065</author><pubDate>Thu, 26 Nov 2020 18:51:59 +0800</pubDate><description><![CDATA[[debug] you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake or configure script in function ‘cvShowImage’
问题描述
you are on Ubuntu or Debian, install libgtk2.0-dev and pkg-config, then re-run cmake or configure script in func]]></description><category></category></item><item><title><![CDATA[[paper reading] FCOS]]></title><link>https://blog.csdn.net/qq_41683065/article/details/109548685</link><guid>https://blog.csdn.net/qq_41683065/article/details/109548685</guid><author>qq_41683065</author><pubDate>Sun, 08 Nov 2020 08:00:06 +0800</pubDate><description><![CDATA[[paper reading] FCOS

本来想放到GitHub的，结果GitHub不支持公式。
没办法只能放到CSDN，但是格式也有些乱
强烈建议去GitHub上下载源文件，来阅读学习！！！这样阅读体验才是最好的
当然，如果有用，希望能给个star！





topic
motivation
technique
key element
math
use yourself
relativity




FCOS
IdeaContribution
FCOS ArchitectureCenter-nessM]]></description><category></category></item><item><title><![CDATA[[paper reading] CenterNet (Object as Points)]]></title><link>https://blog.csdn.net/qq_41683065/article/details/109548508</link><guid>https://blog.csdn.net/qq_41683065/article/details/109548508</guid><author>qq_41683065</author><pubDate>Sat, 07 Nov 2020 16:19:28 +0800</pubDate><description><![CDATA[[paper reading] CenterNet (Object as Points)

本来想放到GitHub的，结果GitHub不支持公式。
没办法只能放到CSDN，但是格式也有些乱
强烈建议去GitHub上下载源文件，来阅读学习！！！这样阅读体验才是最好的
当然，如果有用，希望能给个star！





topic
motivation
technique
key element
math
use yourself
relativity




CenterNet(Object as Points)]]></description><category></category></item><item><title><![CDATA[[paper reading] CenterNet (Triplets)]]></title><link>https://blog.csdn.net/qq_41683065/article/details/109548342</link><guid>https://blog.csdn.net/qq_41683065/article/details/109548342</guid><author>qq_41683065</author><pubDate>Sat, 07 Nov 2020 16:10:43 +0800</pubDate><description><![CDATA[[paper reading] CenterNet (Triplets)

本来想放到GitHub的，结果GitHub不支持公式。
没办法只能放到CSDN，但是格式也有些乱
强烈建议去GitHub上下载源文件，来阅读学习！！！这样阅读体验才是最好的
当然，如果有用，希望能给个star！





topic
motivation
technique
key element
math
use yourself
relativity




CenterNet(triple)
Problem to SolveI]]></description><category></category></item><item><title><![CDATA[[paper reading] CornerNet]]></title><link>https://blog.csdn.net/qq_41683065/article/details/109548057</link><guid>https://blog.csdn.net/qq_41683065/article/details/109548057</guid><author>qq_41683065</author><pubDate>Sat, 07 Nov 2020 16:01:21 +0800</pubDate><description><![CDATA[[paper reading] CornerNet

本来想放到GitHub的，结果GitHub不支持公式。
没办法只能放到CSDN，但是格式也有些乱
强烈建议去GitHub上下载源文件，来阅读学习！！！这样阅读体验才是最好的
当然，如果有用，希望能给个star！





topic
motivation
technique
key element
math
use yourself
relativity




CornerNet
Problem to SolveKeyPoints (Anchor-Fr]]></description><category></category></item><item><title><![CDATA[[paper reading] RetinaNet]]></title><link>https://blog.csdn.net/qq_41683065/article/details/109547868</link><guid>https://blog.csdn.net/qq_41683065/article/details/109547868</guid><author>qq_41683065</author><pubDate>Sat, 07 Nov 2020 15:44:07 +0800</pubDate><description><![CDATA[[paper reading] RetinaNet

本来想放到GitHub的，结果GitHub不支持公式。
没办法只能放到CSDN，但是格式也有些乱
强烈建议去GitHub上下载源文件，来阅读学习！！！这样阅读体验才是最好的
当然，如果有用，希望能给个star！





topic
motivation
technique
key element
math
use yourself
relativity




RetinaNet
Problem to SolveNegative ==&gt; Easy]]></description><category></category></item><item><title><![CDATA[[paper reading] SSD]]></title><link>https://blog.csdn.net/qq_41683065/article/details/109547683</link><guid>https://blog.csdn.net/qq_41683065/article/details/109547683</guid><author>qq_41683065</author><pubDate>Sat, 07 Nov 2020 15:34:10 +0800</pubDate><description><![CDATA[[paper reading] SSD

本来想放到GitHub的，结果GitHub不支持公式。
没办法只能放到CSDN，但是格式也有些乱
强烈建议去GitHub上下载源文件，来阅读学习！！！这样阅读体验才是最好的
当然，如果有用，希望能给个star！





topic
motivation
technique
key element
use yourself
math




SSD
Problem to SolveContributions
SSD ArchitecturePros &amp; Co]]></description><category></category></item></channel></rss>