access

yshep server 2 ip: 165.132.111.50

setup environment in anaconda (ROOT, pytorch, qiskit etc.)

WRARNING!!! this recipe with csh or tcsh based

add the following in your .cshrc (shell setup file => remove '!' in all below lines when you copy to your .cshrc file)

!# >>> conda initialize >>>

!# !! Contents within this block are managed by 'conda init' !!

!if ( -f "/home/anaconda3/etc/profile.d/conda.csh" ) then

! source "/home/anaconda3/etc/profile.d/conda.csh"

!else

! setenv PATH "/home/anaconda3/bin:$PATH"

!endif

!# <<< conda initialize <<<

setup ROOT in anaconda (https://root.cern)

conda create -c conda-forge --name root_torch root

conda activate root_torch

conda config --add channels conda-forge

conda update --all ## do this whenever you install some packages like below

setup pytorch in the activated conda (https://pytorch.org)

conda install pytorch torchvision torchaudio cpuonly -c pytorch

(or try if above doesn't work: conda install pytorch torchvision torchaudio cudatoolkit=11.1 -c pytorch -c nvidia)

install qiskit (for QML with IBM QC resource)

pip install qiskit

pip install 'qiskit-terra[visualization]'

pip install 'qiskit[machine-learning]'

in the case you need "uproot" package

conda install -c conda-forge uproot

how to access Jupiter notebook remotely

in remote server (like yshep server 2)

conda install jupyter

jupyter notebook --generate-config

vi ~/.jupyter/jupyter_notebook_config.py => c.NotebookApp.allow_remote_access=True

jupyter notebook --port=9000 --no-browser & ### kill this job if you want to disconnect with the server

jupyter notebook password ## setup your password

at local terminal

ssh -N -f -L 8000:localhost:9000 -p 4000 (your id)@165.132.111.50

http://localhost:8000/tree? ## in your browser (like chrome, safari etc.)

!! you should click "quit" in your browser to finish it

Test ML example in ROOT

cd $ROOTSYS/tutorials/tmva/ ### in your terminal (in remote server)

root -l ./TMVAClassification.C("BDT,SVM,DNN_CPU,DNN_GPU,MLP") ## run various ML methods, many other options available in the file

Test DNN example in jupyter notebook

! in jupyter notebook, follow ClassificationPyTorch.py

cd $ROOTSYS/tutorials/tmva/pytorch/ ### in your terminal (in remote server)

wget http://root.cern.ch/files/tmva_class_example.root ### in your terminal (in remote server)

! copy ClassificationPyTorch.py to your jupyer notebook and run

! example attached: https://github.com/hdyoo/ML_test/blob/main/Test_Pytorch.ipynb