Optimization for Machine Learning
Time: 09/03/2021 - 18/03/2021 Teachers: Chuan Xu, Othmane Marfoq
This is a Lab course for the main course Optimization for Machine learning
Tutorial for the usage of NEF cluster:
(Please finish the first step before 03/03/2021, finish the rest of the steps before 09/03/2021)
Create a private account for the Inria nef cluster here [ open the hyperlinks, please click twice on them.]
[Please do it as soon as possible before 03/03/2021, as it may take time to create the account]:
for company : please choose "other"
for mail address : please use your UCA email [xxxx@etu.univ-cotedazur.fr](no gmail, etc.)
expiration date : 2021-04-09
ssh public key : generate ssh key (Linux, MacOS, Windows)
and attach the public one (ended by .pub)
reminder - openssh format is needed for the key (windows/putty users will need to convert the key to openssh format - tutorial)
usecase description : please indicate "Optimization for Machine Learning" (Master UCA Data Science and AI)
account creation may require several days, a confirmation email will be sent once the account is created
2. Log in to the cluster once the account is created:
you may try to transfer your own local file from computer to the cluster directory by:
For Linux and MacOS: scp localfile username@nef-frontal.inria.fr:~/
For Windows: pscp localfile username@nef-frontal.inria.fr:~/ [Tutorial]
3. Reserve computing resources in the cluster and log into the reserved node:
oarsub -l /nodes=1/core=2,walltime=1 -I (reserving 2 CPU cores from one node for one hour)
4. Get Pytorch prepared: mutiple ways , recommend Method 3.
To check if Pytorch is well installed, from the terminal type:
[Please make sure the following commands work before the start of the TD, if not, please contact chuan.xu@inria.fr]
python
import torch
print(torch.__version__)
print(torch.version.cuda)
print(torch.backends.cudnn.version())
5. Python editor in cluster, recommend to use VIM and know the basic operation in advance.
6. If you have any question concerning the previous points, please contact me: chuan.xu@inria.fr.
7. Please install Pytorch on your own computer following: https://pytorch.org/get-started/locally
TD1: 09/03/2021 (9h00-12h00)
Colab: Usage of NEF, PyTorch, package for distributed communication
TD2: 11/03/2021 (9h00-12h00)
TD2: 15/03/2021 (14h00-17h00)
TD4: 18/03/2021 (9h00-12h00)