ODU HPC

 Since HPC is generally a type of computer that is familiar to many researchers, we offered "Intro to HPC" workshops to help people onboard to using the system. We just had offered the last one last week. We recorded the workshop—anyone is welcome to watch this to learn:

“Introduction to HPC -- 2025-02-25 Workshop raw recording”

https://odumedia.mediaspace.kaltura.com/media/1_to1ksub3

Length: 03hr 32m 45s

Slides: https://drive.google.com/file/d/1h4-E335c4YnaiClAO9N3A63PSq6YYqXo/view .

Our complete  documentation can be read at: https://wiki.hpc.odu.edu/ .

ODU HPC request form

 

HPC access request form

https://forms.odu.edu/view.php?id=93440

whisperx wahab, warnings

 

[hqin@wahab-01 ~]$ [hqin@wahab-01 ~]$ [hqin@wahab-01 ~]$ module load pytorch-gpu/2.2[hqin@wahab-01 ~]$ crun -p ~/envs/whisperx python Python 3.10.13 (tags/v3.10.13-25-g07fbd8e9251-dirty:07fbd8e9251, Sep 27 2023, 23:32:09) [GCC 13.2.0] :: Intel Corporation on linuxType "help", "copyright", "credits" or "license" for more information.>>> import whisperx/opt/conda/lib/python3.10/site-packages/torch/cuda/__init__.py:628: UserWarning: Can't initialize NVML warnings.warn("Can't initialize NVML")INFO:speechbrain.utils.quirks:Applied quirks (see `speechbrain.utils.quirks`): [disable_jit_profiling, allow_tf32]INFO:speechbrain.utils.quirks:Excluded quirks specified by the `SB_DISABLE_QUIRKS` environment (comma-separated list): []

waha github clone

 on odu waha, github can be used with RSA publication key. 

For example

git clone git@github.com:hongqin/AI4Health.git

ODU HPC request form

 Please direct all students who want to use HPC to fill this form: https://forms.odu.edu/view.php?id=93440 .

 

SLP hpc log

 https://wiki.hpc.odu.edu/Software/Python

https://drive.google.com/file/d/1G8SWxbR-yLtLWy4GGSTds3w3ItrbjPq-/view

module load transformers/4.41

crun -p ~/envs/SLP.new python mycode.py

slurm job. 

[hqin@wahab-01 ~/SLPsummer24]$ cat whi*sh#!/bin/bashecho -n "Run whisher small job "hostnamemodule load transformers/4.41crun -p ~/envs/SLP.new python ~/SLPsummer24/whisper_transcribe_small.py

revised slurm job

[hqin@wahab-01 ~/SLPsummer24]$ cat *sh#!/bin/bash#SBATCH --job-name=smallWhisper # Specify the job name#SBATCH --partition=gpu # Specify the GPU partition#SBATCH --gres=gpu:1 # Request 1 GPU#SBATCH -c 4 # Request 4 CPU cores#SBATCH --time=24:00:00 # Set a time limit for the job#SBATCH --output=hq-whisper-small.out # Output fileecho -n "Run whisher small job "hostnamemodule load transformers/4.41crun -p ~/envs/SLP.new python ~/SLPsummer24/whisper_transcribe_small.py