<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[SPiriTPIX的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/SPiriTPIX</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; SPiriTPIX]]></copyright><item><title><![CDATA[2021-06-21]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/118092650</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/118092650</guid><author>SPiriTPIX</author><pubDate>Mon, 21 Jun 2021 19:00:59 +0800</pubDate><description><![CDATA[import cv2
import numpy as np
import torch





int main()
{
    return 0;
}





]]></description><category></category></item><item><title><![CDATA[【基建笔记】mujoco和mujoco_py安装的坑]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/106376527</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/106376527</guid><author>SPiriTPIX</author><pubDate>Wed, 27 May 2020 16:10:26 +0800</pubDate><description><![CDATA[介绍mujoco和mujoco_py过程和遇到问题解决方法
-1. 吐槽前言:

不少技术文章还是摘抄太多, 大同小异, 千篇一律, 知其然不知其所以然, 所以常常一错百错.
相同问题的解决方案虽都是相似的, 但很多博文都没有讲清楚重点. 而且过程博主自己都未必亲自验证过, 缺乏责任意识.
不少博文给出的方法, 看是大而全面, 有理有据. 但不够细致入微,不能直接利用. 严重影响学习效率.

0. 环境版本说明:

ubuntu-16.04
gym-0.15.4
tensorflow 1.12.0
CUDA]]></description><category></category></item><item><title><![CDATA[【基建笔记】通过本地文件安装 pytorch 和 torchvision]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/105406180</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/105406180</guid><author>SPiriTPIX</author><pubDate>Thu, 09 Apr 2020 11:36:59 +0800</pubDate><description><![CDATA[通过本地文件安装pytorch1.4.0和torchvision0.5.0
0.说明:

用Pytorch官网中提供的安装命令真的巨慢,
即便在conda中添加了清华源路径还是如此. 具体操作可见 参考博客
所以直接在清华源找到对应的版本文件,通过浏览器下载后再手动安装. https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/lin...]]></description><category></category></item><item><title><![CDATA[【基建笔记】深度学习基础工具:CUDA + cuDNN安装与配置]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/105400759</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/105400759</guid><author>SPiriTPIX</author><pubDate>Thu, 09 Apr 2020 11:11:27 +0800</pubDate><description><![CDATA[安装cuda10.0+cudnn7.4+tensorflow-gpu
最终目的是为了安装tensorflow-gpu版本，如果不需要gpu版本的可以直接用pip进行安装即可。否则需要安装nvidia的cuda和cudnn。
cuda-cudnn-tensorflow-gpu版本对应表

0. 卸载cuda老版本

先卸载老版本内容. 我之前用的9.0, 如果上一次是默认安装,整个安装目录应该是/u...]]></description><category></category></item><item><title><![CDATA[【基建笔记】ubuntu中安装terminator并美化]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/103724577</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/103724577</guid><author>SPiriTPIX</author><pubDate>Fri, 27 Dec 2019 01:17:29 +0800</pubDate><description><![CDATA[需求说明

ubuntu下终端是经常使用的工具,但自带的无法直接分屏,使用经典的terminator
terminator自带的主题和风格不堪入目,尝试优化.
当前电脑已安装on_my_zsh,我自己用的比较舒服,美化目标.

解决方案
1. 安装terminator
无需添加任何其他源,直接用sudo apt install terminator安装即可.如果出问题,建议先使用sudo apt ...]]></description><category></category></item><item><title><![CDATA[【工程笔记】ROS-kinetic中安装特定版本的Gazebo]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/103108073</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/103108073</guid><author>SPiriTPIX</author><pubDate>Sun, 17 Nov 2019 14:22:07 +0800</pubDate><description><![CDATA[需求说明

需要Gazebo-8.0
安装ROS时使用sudo apt install ros-kinetic-desktop-full，所以默认安装了Gazebo-7.0

解决思路

卸载gazebo-7.0并安装gazebo-8.0及其所有和ros-kinetic的关联包
ROS和gazebo是两个独立分离的组件，可以分别安装

1. 整理ROS
1.1 如果没有安装ROS则先安装ROS

...]]></description><category></category></item><item><title><![CDATA[【论文笔记】MIT-基于DRL的行人环境移动机器人导航]]></title><link>https://blog.youkuaiyun.com/SPiriTPIX/article/details/99101829</link><guid>https://blog.youkuaiyun.com/SPiriTPIX/article/details/99101829</guid><author>SPiriTPIX</author><pubDate>Sat, 10 Aug 2019 19:21:35 +0800</pubDate><description><![CDATA[MIT_papers_reading




Num
title
author
time




1
Decentralized Non-communicating Multiagent Collision Avoidance with Deep Reinforcement Learning
Yu Fan Chen, Miao Liu, Michael Everett, and Jonathan ...]]></description><category></category></item></channel></rss>