Skip to content

dpclip/dpclip

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commits
 
 
 
 
 
 

Repository files navigation

Training a Differentially Private CLIP

Useful Commands

To create a new tmux session to close your laptop while running: tmux new (can also name your session if you want to access it later).

Use conda to create an environment with all installed packages: conda activate myenv

Request for resources from the FAS cluster: salloc -p gpu_test -t 0-01:00 --mem 8000 --gres=gpu:1 Can specify which gpu to use, e.g. nvidia_a100-sxm4-80gb

OR (Recommended) do srun -p seas_gpu -t 6-10:00 --mem 80000 --gres=gpu:nvidia_a100-sxm4-80gb:1 nohup python3 xxx.py & seas_gpu is optional

To view currently running jobs: squeue

To cancel a job: scancel <job_id>

To run a job: python3 dpclip.py > output.txt &

To direct output to o.txt and process/error to another file e.txt and show the pid: nohup python3 programname.py 1> o.txt 2> e.txt &

To run a job on the i^th gpu: CUDA_VISIBLE_DEVICES=i

To show jobs running on gpus: nvidia-smi

To cancel all jobs: scancel -u username

Fashion MNIST/MNIST benchmarks

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published