Jupyter Server

Running jupyter

jupyter notebook --no-browser --port=1717
jupyter lab --port=1717

Connecting to server

ssh -L 1717:localhost:1717 [email protected]

Using Conda is a better alternative

Don’t forget to add conda-forge channel

conda create -n venv python=3.8

The python part is necessary! If you don’t choose that, no python would be installed!

SSH-Copy-id

First make sure you have an ssh key in your computer. Then you have to copy it to server using following command:

ssh-copy-id -i ~/.ssh/id_rsa.pub [email protected]

Also, you can first set .ssh/config file. Then use following command.

ssh-copy-id servername

nvidia-smi

pip install nvidia-htop
pip install nvitop
 
nvidia-htop.py
nvitop

nvidia-smi --query (--display)
nvidia-smi -q
nvidia-smi -i 0
nvidia-smi -i 0 -q -d "memory"
nvidia-smi --query-gpu='memory.free' --format=csv

https://codeyarns.com/tech/2019-04-26-nvidia-smi-cheatsheet.html

https://briot-jerome.developpez.com/fichiers/blog/nvidia-smi/list.txt

Connect to Another GPU

torch.device(‘cuda:0’) or torch.device(‘cuda:1’)

export CUDA_VISIBLE_DEVICES=1

CUDA_VISIBLE_DEVICES=1 ./cuda_executable

Not needed

Creating venv

virtualenv -p python3 venv source venv/bin/activate

install multiple python kernels with multiple versions

Install pyenv $c u r l [h ttp s : // p y e n v . r u n] (h ttp s : // p y e n v . r u n /) ∣ ba s h F o l l o w t h e in s t r u c t i o n t o m o d i f y /. ba s h r c I n s t a l l t h e l a t es tP y t h o n f r o m so u r ceco d e$ pyenv install 3.8.3 Check installed Python versions $p y e n v v er s i o n s S w i t c h P y t h o n v er s i o n$ pyenv global 3.8.3 Check where Python is actually installed $p y e n v p r e f i x / h o m e / a d min /. p y e n v / v er s i o n s /3.8.3 C h ec k t h ec u r r e n tP y t h o n v er s i o n$ python -V Python 3.8.3

Installing Kernel for jupyter (only once)

python -m venv myenv source myenv/bin/activate

pip install ipykernel python -m ipykernel install —user —name=myenv

after having pipenv

is this necessary? or even good?

You can switch to environment using pipenv shell see pipenv current environment setting pipenv —venv change pipenv current virtual environment echo ~/myenv > .venv

Python linux Python Libraries for Training

نوشته‌های امیر

Explorer

Python Stuff to install on while working on server