【jetson nano】 aarch64安装pycharm pytorch cuda

柔光的暖阳◎ 2022-11-12 05:28 112阅读 0赞

sudo vi ~/.bashrc 编辑环境变量  
source ./profile 编译  
sudo mv jdk1.8.0\_281 /usr/local/java/

先记录过程，目前还有问题待解决。

export PATH=/usr/local/cuda-10.2/bin:$PATH
    export LD_LIBRARY_PATH=/usr/local/cuda-10.2/lib64\
    >                          ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

nvcc --version  
![在这里插入图片描述][20210323124018606.png]  
./pycharm.sh  
./pycharm.sh: 218: ./pycharm.sh: /usr/local/java/jdk1.8.0\_281/bin/java: Exec format error  
java -version  
bash: /usr/local/java/jdk1.8.0\_281/bin/java: 无法执行二进制文件: 可执行文件格式错误  
sudo dpkg -i baidunetdisk\_3.5.0\_amd64.deb  
\[sudo\] he 的密码：  
dpkg: 处理归档 baidunetdisk\_3.5.0\_amd64.deb (–install)时出错：  
软件包体系架构(amd64)与本机系统体系结构(arm64)不符  
在处理时有错误发生：  
baidunetdisk\_3.5.0\_amd64.deb

sudo dpkg --add-architecture amd64
    sudo dpkg -i baidunetdisk_3.5.0_amd64.deb

dpkg: 处理软件包 baidunetdisk:amd64 (–install)时出错：  
依赖关系问题 - 仍未被配置  
正在处理用于 gnome-menus (3.13.3-11ubuntu1.1) 的触发器 …  
正在处理用于 desktop-file-utils (0.23-1ubuntu3.18.04.2) 的触发器 …  
正在处理用于 bamfdaemon (0.5.3+18.04.20180207.2-0ubuntu1) 的触发器 …  
Rebuilding /usr/share/applications/bamf-2.index…  
正在处理用于 mime-support (3.60ubuntu1) 的触发器 …  
正在处理用于 hicolor-icon-theme (0.17-2) 的触发器 …  
在处理时有错误发生：  
baidunetdisk:amd64

pip3 install torch-1.5.0-cp36-cp36m-linux\_aarch64.whl  
Processing ./torch-1.5.0-cp36-cp36m-linux\_aarch64.whl

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMzU4NTc0_size_16_color_FFFFFF_t_70]  
wget https://nvidia.box.com/shared/static/veo87trfaawj5pfwuqvhl6mzc5b55fbj.whl -O torch-1.1.0a0+b457266-cp36-cp36m-linux\_aarch64.whl

he@he-desktop:~/Desktop$ pip3 install torch-1.5.0-cp36-cp36m-linux\_aarch64.whl  
Processing ./torch-1.5.0-cp36-cp36m-linux\_aarch64.whl  
Exception:  
Traceback (most recent call last):  
File “/usr/lib/python3/dist-packages/pip/basecommand.py”, line 215, in main  
status = self.run(options, args)  
File “/usr/lib/python3/dist-packages/pip/commands/install.py”, line 353, in run  
wb.build(autobuilding=True)  
File “/usr/lib/python3/dist-packages/pip/wheel.py”, line 749, in build  
self.requirement\_set.prepare\_files(self.finder)  
File “/usr/lib/python3/dist-packages/pip/req/req\_set.py”, line 380, in prepare\_files  
ignore\_dependencies=self.ignore\_dependencies))  
File “/usr/lib/python3/dist-packages/pip/req/req\_set.py”, line 620, in \_prepare\_file  
session=self.session, hashes=hashes)  
File “/usr/lib/python3/dist-packages/pip/download.py”, line 809, in unpack\_url  
unpack\_file\_url(link, location, download\_dir, hashes=hashes)  
File “/usr/lib/python3/dist-packages/pip/download.py”, line 715, in unpack\_file\_url  
unpack\_file(from\_path, location, content\_type, link)  
File “/usr/lib/python3/dist-packages/pip/utils/**init**.py”, line 617, in unpack\_file  
flatten=not filename.endswith(’.whl’)  
File “/usr/lib/python3/dist-packages/pip/utils/**init**.py”, line 506, in unzip\_file  
data = zip.read(name)  
File “/usr/lib/python3.6/zipfile.py”, line 1338, in read  
return fp.read()  
File “/usr/lib/python3.6/zipfile.py”, line 858, in read  
buf += self.\_read1(self.MAX\_N)  
File “/usr/lib/python3.6/zipfile.py”, line 948, in \_read1  
data = self.\_decompressor.decompress(data, n)  
MemoryError

更新：  
通过在32G内存卡上一系列失败操作，最大的感触是想跑深度学习一定不要用32G！！！！买来64G之后烧录官方带有环境的镜像，pytorch什么的很快就弄好了！！然后python可以在jupyter里面编辑。总之，64G是对低限度。感觉128G更好…

[20210323124018606.png]: /images/20221022/e0257b29ed1449d3aabbb38b1be9ab3d.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQxMzU4NTc0_size_16_color_FFFFFF_t_70]: /images/20221022/ec3593d5040e41aeb27b9c2167c30d87.png