Ubuntu 18.04 Environmental settings

Table of Contents

Basic setting

• Ubuntu installation
• Network connection
• Software updater
• Graphic card driver
• Language setting
• Anaconda

Tensorflow

• CUDA
• CUDNN
• Tensorflow

Pytorch

• Pytorch

ETC

• Pycharm
• Pytorch tutorial
• Other programs
• Useful command
• Teamviewer
• KakaoTalk
• Wheel speed

Useful command

Ubuntu installation

Version : 18.04

[reference]

• Download ubuntu 18.04 [Link]
• Prepare USB
• Create booting disk [Link(Korean)]
• Turn off Windows Quick Start [Link(Korean)]
• BIOS
  • Boot Ubuntu USB
• Install Ubuntu
  • Welcome : English
  • Keyboard layout : ENG(US)
  • Updates and other software : Normal installation / Download updates while installing Ubuntu
  • Installation type : something else
    • I installed it on the remaining hard disk.
    • And, I partitioned it into 'partition swap' and 'ext4'. [Link]
  • Where are you : seoul
  • Who are you : write your information

---

Network connection

• Settings
• Network
• Wired option
• IPv4
  • Input (Address, Netmask, Gateway, DNS)

---

Software updater

• Searching window
• software update

---

Graphic card driver

[reference]

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo ubuntu-drivers autoinstall
sudo reboot

(after reboot)

sudo nvidia-settings

---

Language setting

(Korean only)

[reference]

• Settings
• Region & Language
• Input sources
  • Add (+ button) Korean(Hangul) <- If 'Korean(Hangul)' (not 'Korean') does not exist, click 'Manage Installed Languages' and install it
  • Delete (- button) English

(Optional - shortcut setting)

• Input sources
• Click Korean(Hangul)
• Option button (It exists only in 'Korean(Hangul)' not ' Korean')
• Hangul toggle key
  • Add HAN/ENG key (Alt-R or Hangul)

---

Anaconda

Version : Python 3.7 version (64bit)

[reference]

• Download anaconda (click the reference)

cd
cd Downloads/
bash Anaconda3-5.3.0-Linux-x86_64.sh

• License agreement
• Confirm install location
• /root/.bachrc? [yes]
• VSCode? [No]

(after reboot)

conda --version

---

CUDA

Version 9.0

[reference]

• Install CUDA 9.0 [Link]
  • Linux / x86_64 / Ubuntu / 17.04 (18.04 is not supported) / runfile (local)
  • Base Installer (Download 1.6GB)

cd
cd Downloads/
sudo chmod +x cuda_9.0.176_384.81_linux.run
./cuda_9.0.176_384.81_linux.run --override

• EULA? [accept]
• Unsupported configuration? [yes]
• Graphic driver? [no]
• Cuda toolkit? [yes]
• Confirm toolkit location
• run with 'sudo'? [yes]
• symbolic link? [no]
• Cuda samples? [no]
• Confirm sample location

nvcc --version

Version 10.2

• Install CUDA 10.2 [Link]
  • wget http://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run
  • sudo sh cuda_10.2.89_440.33.01_linux.run
  • Graphic driver? [no] (important) / symbolic link ? [no] (directly connect individual version by bashrc)

(Optional - install multiple CUDA versions)

• When installing one version according to the above procedure and installing another one, the following error message may occur. (ex: Ubuntu 18.04 + CUDA8.0)

# Command lines
===========
= Summary =
===========
Driver: Not Selected
Toolkit: Installation Failed
Samples: Installation Failed

Logfile is /tmp/cuda_install_13486.log
Signal caught, cleaning up

# Log file
Uncompressing NVIDIA CUDA....................................................................................

Can't locate InstallUtils.pm in @INC (you may need to install the InstallUtils module) (@INC contains: /etc/perl /usr/local/lib/x86_64-linux-gnu/perl/5.26.1 /usr/local/share/perl/5.26.1 /usr/lib/x86_64-linux-gnu/perl5/5.26 /usr/share/perl5 /usr/lib/x86_64-linux-gnu/perl/5.26 /usr/share/perl/5.26 /usr/local/lib/site_perl /usr/lib/x86_64-linux-gnu/perl-base) at ./install-linux.pl line 6.

BEGIN failed--compilation aborted at ./install-linux.pl line 6.

Verifying archive integrity... All good.

Uncompressing NVIDIA CUDA Samples.......................................................................................

Can't locate InstallUtils.pm in @INC (you may need to install the InstallUtils module) (@INC contains: /etc/perl /usr/local/lib/x86_64-linux-gnu/perl/5.26.1 /usr/local/share/perl/5.26.1 /usr/lib/x86_64-linux-gnu/perl5/5.26 /usr/share/perl5 /usr/lib/x86_64-linux-gnu/perl/5.26 /usr/share/perl/5.26 /usr/local/lib/site_perl /usr/lib/x86_64-linux-gnu/perl-base) at ./install-sdk-linux.pl line 6.

BEGIN failed--compilation aborted at ./install-sdk-linux.pl line 6.
'uninstall_cuda_8.0.pl' -> '/usr/local/cuda-8.0/bin/uninstall_cuda_8.0.pl'

• Type the following command and reinstall the CUDA.
  • Unpack .run file ./cuda*.run --tar mxvf
  • Copy InstallUtils.pm file cp InstallUtils.pm /usr/lib/x86_64-linux-gnu/perl-base
  • export $PERL5LIB

(optional - install CUDA 10.0 + CUDNN 7.5)

---

CUDNN

Version 7.0.5

[reference]

• Download this “cuDNN v7.0.5 Library for Linux”
• Unpack the archive and move it the contents into the directory where you install CUDA 9.0

# Unpack the archive
tar -zxvf cudnn-9.0-linux-x64-v7.tgz
# Move the unpacked contents to your CUDA directory
sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda-9.0/lib64/
sudo cp  cuda/include/cudnn.h /usr/local/cuda-9.0/include/
# Give read access to all users
sudo chmod a+r /usr/local/cuda-9.0/include/cudnn.h /usr/local/cuda-9.0/lib64/libcudnn*

• Install libcupti

sudo apt-get install libcupti-dev

• Do the CUDA post-install actions

gedit ~/.bashrc

• Write the below commands

export PATH="/usr/local/cuda-9.0/bin:$PATH"
export LD_LIBRARY_PATH="/usr/local/cuda-9.0/lib64:$LD_LIBRARY_PATH"

• Restart ! or source ~/.bashrc

(Optional - Other version)

• Other version [reference]

(Optional - CUDNN PATH)

• Download CUDNN and UNPACK
• move "lib64 and include folders" to /home/$pID/cudnn/$version
• Add the below commands export LD_LIBRARY_PATH="/home/$pID/cudnn/$version/lib64:$LD_LIBRARY_PATH"

---

Tensorflow

[reference]

• Create virtual environment for Tensorflow by Anaconda

conda create -n py36_tensorflow python=3.6
conda activate py36_tensorflow
conda deactivate

• Install tensorflow (latest version)

conda activate py36_tensorflow
pip install --upgrade tensorflow-gpu
python -c "import tensorflow as tf; print(tf.__version__)"

---

• Download other versions

# example
pip install tensorflow-gpu==1.4.0

Pytorch

[reference]

• Create virtual environment for Pytorch by Anaconda
• Install pytorch (various versions)

conda create -n seokeon_py36_torch041 python=3.6
source activate seokeon_py36_torch041
conda install pytorch=0.4.1 cuda90 -c pytorch
source deactivate

conda create -n seokeon_py27_torch041 python=2.7
source activate seokeon_py27_torch041
conda install pytorch=0.4.1 cuda90 -c pytorch
source deactivate

conda create -n seokeon_py36_torch031 python=3.6
source activate seokeon_py36_torch031
conda install pytorch=0.3.1 cuda90 -c pytorch
source deactivate

conda create -n seokeon_py27_torch031 python=2.7
source activate seokeon_py27_torch031
conda install pytorch=0.3.1 cuda90 -c pytorch
source deactivate

• Install pytorch (latest version)

conda install pytorch-cpu torchvision-cpu -c pytorch
conda install pytorch torchvision cudatoolkit=8.0 -c pytorch
conda install pytorch torchvision cudatoolkit=9.0 -c pytorch
conda install pytorch torchvision cudatoolkit=10.0 -c pytorch

---

Pycharm

[reference]

• Install pycharm using Snaps

sudo snap install pycharm-community --classic 
# or sudo snap install pycharm-professional --classic

• pop up the message "pycharm-community 2017.3.3 from 'jetbrains' installed"

pycharm-community

• Make new project

• Environmental settings

  • File
  • Settings
  • Project interpreter
  • Select the virtual environment where you installed pytorch

• Change pycharm keymap

  • File
  • Settings
  • Keymap
    • comment : ctrl+R
    • run : F5
    • debug : F6
    • resume program : F7
    • close (Editor Tabs) : ctrl+W
    • Quick Evaluate Expression : Shift+F8
    • Evaluate Expression : F8
    • step over : F10
    • step into : F11
    • step out : shift + F11
    • Toggle line breakpoint : F12
    • Code->Folding->Collapse All: Ctrl+Alt+minus

• Run/Debug configurations

  • Python interpreter (python directory in the virtual environment)
  • Environment variables (optional.. for TensorFlow)
    • Add (Name : LD_LIBRARY_PATH / Value : /usr/local/cuda-9.0/lib64)
  • Working directory
    • /home/user_name/PycharmpProjects/your_project/

---

Pytorch tutorial

[reference]

Other programs

• In the vitual environment (anaconda)

conda activate pytorch36
pip install scipy
pip install sacred
pip install matplotlib
pip install opencv-python
pip install pillow
pip install numpy
conda install -c pytorch torchvision
conda install -c anaconda pyyaml
conda deactivate

Teamviewer

[reference]

KakaoTalk

[reference]

• Caution! An error may occur

• Install Wine and environmental settings

sudo apt install wine-stable cabextract
WINEARCH=win32 WINEPREFIX=~/.wine wine wineboot
wget  https://raw.githubusercontent.com/Winetricks/winetricks/master/src/winetricks
chmod +x winetricks
./winetricks --optout

• Select the default winprefix

• Install a Windows DLL or component

• Select (gdiplus, riched30, wmp9, msxml6)

• Copy Gulim font (window to ubuntu)

• copy C:/Windows/Fonts/gulim.ttf (or ttc) -> ~/.wine/drive_c/windows/Fonts (using cp -i or something else)

chmod 644 ~/.wine/drive_c/windows/Fonts/gulim.ttf
gedit ~/.wine/system.reg

Change from

"MS Shell Dlg"="Tahoma"
"MS Shell Dlg 2"="Tahoma"

to

"MS Shell Dlg"="Gulim"
"MS Shell Dlg 2"="Gulim"

• Install kakaoTalk

winecfg

• Confirm the version of window
• Download the file (version check) [Download]
• Install it

wine-stable KakaoTalk_Setup.exe

• Change language setting

gedit ~/.local/share/applications/wine/Programs/KakaoTalk/KakaoTalk.desktop

Change from

Exec=env WINEPREFIX="/home/ubuntu/.wine" wine-stable C:\\\\windows\\\\command\\\\start.exe /Unix /home/ubuntu/.wine/dosdevices/c:/ProgramData/Microsoft/Windows/Start\\ Menu/Programs/KakaoTalk/KakaoTalk.lnk

to

Exec=env WINEPREFIX="/home/ubuntu/.wine" LANG="ko_KR.UTF-8" wine-stable C:\\\\windows\\\\command\\\\start.exe /Unix /home/ubuntu/.wine/dosdevices/c:/ProgramData/Microsoft/Windows/Start\\ Menu/Programs/KakaoTalk/KakaoTalk.lnk

• Notably, LANG="ko_KR.UTF-8" is Only added

• Solving the problem of broken font

cd "/home/ubuntu/.wine/dosdevices/c:/Program Files/Kakao/KakaoTalk"
LANG="ko_KR.UTF-8" wine-stable KakaoTalk.exe

• Setting system tray

sudo apt install gnome-shell-extension-top-icons-plus

• Extensions
  • Topicons plus (check!)

Wheel speed

[reference]

sudo apt-get install imwheel
imwheel
sudo gedit /etc/X11/imwheel/startup.conf

• Change ‘IMWHEEL_START=0’ to ‘IMWHEEL_START=1’

gedit ~/.imwheelrc

• Copy all the contents in [ref]

• Add below commands (the number 3 means wheel speed)

".*"
None,      Up,   Button4, 3 
None,      Down, Button5, 3

• End

imwheel -k

Useful command

• compress files : zip -r zipname.zip filename
• unzip : tar -xvzf "file name"
• unzip all 'zip' files :

for file in `ls *.zip`; do unzip "${file}" -d "${file:0:-4}"; done
for file in `ls *.zip`; do unzip "${file}" -d "./"; done
for file in `ls *.rar`; do unrar e "${file}"; done

• zip by 7z:

  • sudo apt-get install p7zip-full
  • 7z a data.7z data.txt (zip)
  • 7z x data.7z (unzip, )

• remove folder : rm -rf "folder name"

• remove 해당 디렉토리 내의 특정파일 삭제: find . -type f -name "*.zip" -exec rm {} ; link

• make folder : mkdir "folder name"

• copy folder : cp -r "folder a" "folder b"

• copy folder (w/o overwrite) : rsync -a -v --ignore-existing src dst

• move folder : mv "folder a" "folder b"

• list subdirectories : tree -d -L 1 find src -mindepth 2 -maxdepth 3 -type d > list.txt

• Visualize gpu situation (auto update)

nvidia-smi -l 1

• Caution! Removing "Alt" function from "Han/Eng" key on the keyboard. [reference]

xmodmap -e 'remove mod1 = Alt_R'
xmodmap -e 'keycode 108 = Hangul'

Option -> Region & Language -> Korean (Hangul) -> Option -> Shortkey (Alt+R -> Hangul)

• Chrome auto scroll [reference]

• Symbolic (soft) link : ln -s target_path(old) link_path(new)

  • ln -s ../../../DB/reid/old_DB ./ 하면 ./ 위치에 old_DB 라는 폴더 생성
  • ln -sf ../../../DB/reid/old_DB ./new_DB 하면 ./ 위치에 new_DB 라는 폴더 생성

• Count the number of files in the certain path find /path/to -type f | wc -l

• Count the number of files in the present path find . -type f | wc -l

• Remove conda env conda env remove -n ENV_NAME

• Personal PATH export PPATH="/path/to" in ~/.bashrc

• Add 'new document' option when right clicking touch ~/Templates/Empty\ Document

• Change ":" to "," in filename find . -name "*:*" -exec rename 's|:|,|g' {} \;

  • sudo apt install rename

• GPU temperature

  • nvidia-smi -q -d temperature -l 1

• 우분투 APT repository 제거하기 [ref]

  • sudo add-apt-repository --remove ppa:~~~~~ (지우길 원하는 프로그램 이름명)

• 터미널 열었을 때 (base) 있는 경우

  • conda config --show | grep auto_activate_base
  • conda config --set auto_activate_base False

• 디스크 용량 체크

  • df -h

• 폴더 내의 용량 체크

  • du -h

• 빠른 삭제

  • sudo rm -r -f /path/

• Compression

  • (install) sudo apt-get install p7zip-full
  • (compress) 7z a /created_file_name/ /folder_name or */
  • (check) 7z l /7z_file/
  • (extract) 7z e /7z_file/

• Memory check

  • watch -d free / watch -n 1 free

• 파일 수 세기

  • find . -type f | wc -l

• 해당 조건 파일 옮기기

  • find path_A -name '*.jpg' -exec mv -t path_B {} +
  • maxdepth 1

• Pycharm deployment (다른 컴퓨터의 pycharm 에서 코드 돌리는 법)

  • original code git commit (or copy your code to new com)
  • git clone repo (or paste the code)
  • Settings -> Build, Execution, Deployment -> new connection (+)
    • Connection: SFTP -> SSH Configuration (make new IP) -> Test connection
    • Mappings: Local path (new com), Deployment path (server com) => same folder
    • If failed at Test connection, in server com or apply VPN (university)
      • sudo apt update -y
      • sudo apt-get install openssh-server
      • sudo service ssh start
      • sudo service openssh-server start
  • Project -> Python Interpreter -> add (+) -> SSH Interpreter -> Existing server configuration -> Interpreter (server -> anaconda -> env -> bin -> python) -> sudo -> Sync folder (setting remote path) -> check automatically upload
  • Setting configuration

• pip 이용한 설치중 Cannot uninstall '~~~' 에러발생

  • sudo pip install pwntools 대신에 sudo pip install --ignore-installed pwntools

• ppt FHD 동영상 저장 방법 (https://m.blog.naver.com/PostView.nhn?blogId=radiobj5&logNo=220345624061&proxyReferer=https:%2F%2Fwww.google.com%2F)

  • 동영상 녹음 후에
  • Alt + F11
  • 삽입 -> 모듈

Sub MkVideo()
    If ActivePresentation.CreateVideoStatus <> ppMediaTaskStatusInProgress Then
    ActivePresentation.CreateVideo FileName:=Environ("USERPROFILE") & "\Desktop\test.mp4", _
    UseTimingsAndNarrations:=True, _
    VertResolution:=1080, _
    FramesPerSecond:=25, _
    Quality:=100
    Else:
    MsgBox "There is another conversion to video in progress"
    End If
    End Sub

• F5

• server 관리

  • ssh 명령어 short-cut관리: gedit ~/.ssh/config
  • ssh 바로 접근: ssh username@ip_adress
  • id 생성
    • 서버 root계정으로 로그인
    • sudo adduser id_name
  • CUDA 설정
    • vim ~/.bashrc
      • Cuda path
        • export PATH="/usr/local/cuda-10.0/bin:$PATH"
        • export LD_LIBRARY_PATH="/usr/local/cuda-10.0/lib64:$LD_LIBRARY_PATH"
      • Anaconda path
        • export PATH="/home/ROOT_NAME/anaconda3/bin:$PATH"
        • export LD_LIBRARY_PATH="/home/ROOT_NAME/anaconda3/lib64:$LD_LIBRARY_PATH"
    • source ~/.bashrc
  • error control
    • anaconda 가상환경 설정시 permission error: sudo chmod -R 777 anaconda3
    • slurm에서 sbatch 안먹을때 (sinfo 입력했을때 drain인경우)
      • 돌아가는 job 있을때: scontrol update nodename=node10 state=resume
      • 돌아가는 job 없을때: scontrol update nodename=node10 state=idle
  • 서버에서 다른 cuda version 쓰고 싶을 때 (10.2 기준)
    • CUDA 파일 다운로드: wget http://developer.download.nvidia.com/compute/cuda/10.2/Prod/local_installers/cuda_10.2.89_440.33.01_linux.run
    • CUDA 설치: sudo sh cuda_10.2.89_440.33.01_linux.run
      • EULA? [accept]
      • Unsupported configuration? [yes]
      • Graphic driver? [no] (important)
      • Cuda toolkit? [yes]
      • run with 'sudo'? [yes]
      • symbolic link? [no] (directly connect individual version by bashrc)
      • Cuda samples? [no]
    • CUDNN 파일 다운로드
    • CUDNN 설치
    • CUDNN 이동
      • move "lib64 and include folders" to /home/choi/cudnn/$version
    • link
      • vim ~/.bashrc
        • export PATH="/usr/local/cuda-10.2/bin:$PATH"
        • export LD_LIBRARY_PATH="/usr/local/cuda-9.0/lib64:$LD_LIBRARY_PATH"
        • export LD_LIBRARY_PATH="/home/choi/cudnn/$version/lib64:$LD_LIBRARY_PATH"
      • source ~/.bashrc
  • 남은 용량 확인 (현재 폴더에서 각각 폴더가 차지하고 있는 양 확인)
    • du -sh *

• conda activate 가 안되고 source activate만 되는 경우

  • source ~/anaconda3/etc/profile.d/conda.sh

• nvidia graphic driver 다른 버전 쓰고 싶을 때 (https://codechacha.com/ko/install-nvidia-driver-ubuntu/)

  • 자동 버전 설치
    • sudo add-apt-repository ppa:graphics-drivers/ppa
    • sudo apt update
    • sudo ubuntu-drivers autoinstall
    • sudo reboot
  • 기존 삭제(만약 기존 설치된 프로그램과 출동한다면): sudo apt --purge autoremove nvidia*

• Automatic mixed precision 쓰는 법 (https://hoya012.github.io/blog/Image-Classification-with-Mixed-Precision-Training-PyTorch-Tutorial/)

  • CUDA10.1 버전 이상만 pytorch1.6 지원
  • CUDA10.1 버전 이상 설치 / CUDNN 설치 (필수인지는 확실하지 않음)
  • conda create -n seokeon_torch16 python=3.6
  • conda activate seokeon_torch16
  • conda install pytorch torchvision cudatoolkit=10.1 -c pytorch (10.2도 가능)
  • nvidia graphic driver 업그레이드
  • 코드에서 변경해야 하는 부분
    • AMP_flag = True
    • if AMP_flag:
      • self.scaler = torch.cuda.amp.GradScaler() [추가된부분]
    • dataloader iteration 내부에서
      • if AMP_flag:
        • with torch.cuda.amp.autocast(): [추가된부분]
          • outputs = self.model(inputs)
          • loss = self.criterion(outputs, labels)
          • self.optimizer.zero_grad()
          • self.scaler.scale(loss).backward() [변경된부분]
          • self.scaler.step(self.optimizer) [변경된부분]
          • self.scaler.update() [변경된부분]
      • else:
        • outputs = self.model(inputs)
        • loss = self.criterion(outputs, labels)
        • self.optimizer.zero_grad()
        • loss.backward()
        • self.optimizer.step()

• Mount 완련!!

  • bootloader가 켜지지 않고 grup gnu terminal 창만 나오는 경우ㅜ
    • ubuntu booting USB로 부팅(try ubuntu without installing)
    • 인터넷연결
    • sudo add-apt-repository ppa:yannubuntu/boot-repair
    • sudo apt-get update
    • sudo apt-get install -y boot-repair
    • boot-repair
    • Click Recommended repair
  • gpt to mbr (https://www.linuxtopic.com/2017/02/convert-partition-table-gpt-to-mbr-in.html)
    • install gdisk
    • gdisk /dev/sda
    • command: r
    • Recovery/transformation command? g
    • (MBR command: p)
    • MBR command: w
    • coverted 1 paritions. Finalize and exit? (Y/N): y
    • (command: w)
    • reboot
  • 4TB이상 하드를 사서 리눅스를 설치할꺼면? (http://blog.naver.com/PostView.nhn?blogId=5bpa&logNo=220460531819)
  • 하드디스크 처음 마운트 (https://seongkyun.github.io/others/2019/03/05/hdd_mnt/)
    • sudo fdisk -l 에서 하드 확인
    • 용량이 2TB 이하인 경우
      • sudo fdisk /dev/sda
      • command: n
      • select: p
      • Partition number: 1
      • First sector: (enter)
      • Last sector: (enter) -> created a new partition ~~
      • command: p
      • command: w
    • format
      • sudo mkfs.ext4 /dev/sda1
    • uuid 확인
      • 해당 disk의 UUID 복사
    • mount
      • sudo mkdir /mnt/directory-to-mount
      • sudo vim /etc/fstab
        • UUID=~~~~~ /directory-to-mount ext4 defaults 0 0
        • 맨 아랫줄에 입력
      • sudo mount -a
      • df -h (마운트 확인)
    • symbolic link
      • sudo ln -s /directory-to-mount /home/choi/
      • cd ~/directory-to-mount
      • sudh chmod 777 ~/directory-to-mount
  • Change mount position

    lsblk # check disk position
    sudo xdg-open /etc/fstab # change disk position
    

    Add /dev/sdc /mnt/hard1 ntfs-3g defaults 0 2 (??? not completed)

  • Unrecognized mount option "default"
    • vim /etc/fstab
    • 에서 default라고 적힌것 defaults로
  • hard가 read-only 인경우 (window caches 에 의해서)
    • https://askubuntu.com/questions/462381/cant-mount-ntfs-drive-the-disk-contains-an-unclean-file-system
    • https://hidekuma.github.io/ec2/ubuntu/linux/remount-ebs/
    • $ cat /proc/mounts 에서 해당 디스크 ro로 되어있는지 확인
    • $ sudo umount /dev/sdXY
    • $ sudo ntfsfix /dev/sdXY (두번?)
    • $ sudo mount -o rw /dev/sdXY

• 7zip 압축

  • 7z a kernel.7z kernel/ -v50m
  • 50mb 분할 압축
  • kernel.7z.001, kernel.7z.002 파일 생성

• 7zip 해제

  • 7z x kernel.7z.001 -aoa

출처: https://ysh0222.tistory.com/26 [Sangho Yoon]

출처: https://ysh0222.tistory.com/26 [Sangho Yoon]

• 다른 서버 폴더 접근

  • 폴더 gui에서
  • connect to server
    • sftp://ID@ip

• 가상환경이나 현재 python에 pip으로 설치된 패키지 목력정보 만들기

  • pip freeze > requirement.txt (문서생성)
  • pip install -r requirements.txt (pip install)

• Git 관련

  • 명령어로 연동
    • git clone ~~~
    • 파일 수정
    • git add --all
    • git commit -m "Fix ~~ or Update ~~"
    • git push origin master
  • pycharm 과 연동 link
    • Web에서 repository 생성
    • VCS>Get from Version Control
    • Github ID login
    • Repository 연결해서 원하는 폴더에 다운로드
    • 원하는 파일 옮겨닮기 (외부에서 옮기면 따로 pycharm에서 add해야하므로 pycharm의 프로젝트 창으로 파일 바로 옮겨줌)
    • commit
    • push
  • 다른 github repository에 pull request를 하는 방법
    • Fork the repository
    • 1. local 작업
      • git clone repository
      • cd repository
      • Create a new branch
        • git branch new-branch
        • git checkout new-branch
        • or
        • git checkout -b new-branch
        • If you want to switch back to master
        • git checkout master
      • Make change locally
        • Modify an existing file or add a new file
        • git add filename.md or git add -A
        • git commit -m "Fixed documentation typos" or git config --global core.editor "nano" (장문을 쓰길 원하면, nano대신 vim가능)
        • git status (생략가능, verify 단계)
        • git push --set-upstream origin new-branch (forked repository가 변경되어 있을 것이다)
    • 2. In repositoy
      • 원하는 파일 수정
    • 3. Pull request (1 or 2 수행 이후)
      • Repository-> Pull request->New pull request

• conda env 복사 붙여넣기 link

  • anaconda version이 다르면 에러날수도
  • conda activate 이름
  • conda env export > environment.yaml
  • python --version (Python 3.6.6) 인 경우
  • conda create --name [이름] python=3.6 (environment.yml 에 있는 이름과 동일, 복사 하고자 하는 서버에 같은 파이선 버전 생성)
  • conda activate [이름]
  • conda env create environment.yml
  • conda 생성 안하고 바로할경우
    • conda env create --prefix <your_conda_env_path> -f environment.yml (envs/이름) 까지