This workshop will give participants a hands-on introduction to using HCC resources for running Artificial intelligence (AI) and Machine Learning (ML) workflows. Participants will learn how to access and utilize GPU resources on HCC's Swan and utilize PyTorch to create and train a simple model. The participants will be introduced to the National Research Platform (NRP), a free research resource for running GPU-based AI and ML workflows.
When: Tuesday, March 25th (UNLincoln) and April 8th (UNLincoln), 12pm to 4:30pm
Where:
- March 25th: UNL City Campus, Jorgensen Hall Room 251
- April 8th: UNL City Campus, Jorgensen Hall Room 251
Cost: FREE!!! Space is limited.
- NOTE: Registration is binding!
Workshop Level: Intermediate.
Required Prerequisites:
- Familiarity with Linux and the command-line
- Experience with Python
- Active HCC account with DUO setup
Learning Objectives:
- GPU access on HCC resources
- Accessing GPU resources on Swan
- Accessing JupyterLab on Swan
- Running AI and ML workflows on Swan
- Using PyTorch
- GPU access on other National resources
Registration:
Attendees will need an HCC account as this workshop will feature hands-on practice.
Schedule:
12pm - 12:15pm |
Setup and SupportTime available for setting up accounts or troubleshooting issues prior to the day's events. |
12:15pm - 1 pm |
Introduction to GPU resources on HCCGeneral introduction to GPU resources Swan offers and its usage. |
1pm - 1:15pm |
Break |
1:15pm - 2pm |
AI and ML workflowsLearn about how to run AI and ML workflows on Swan. |
2pm - 2:10pm |
Break |
2:10pm - 2:50pm |
Introduction to PyTorchLearn how to run GPU PyTorch code on Swan. |
2:50pm - 3pm |
Break |
3pm - 3:50pm |
Introduction to PyTorch (cont.)Learn how to run GPU PyTorch code on Swan. |
3:50pm - 4pm |
Break |
4pm - 4:15pm |
Introduction to National Research Platform (NRP)Learn about different GPU provider. |
4:15pm - 4:30pm |
Open QuestionsTime available for questions about today's events. |
Setup for Hands-On
- Login to Swan's Open OnDemand (OOD) Portal.
- Open a terminal in OOD. Done by opening the "Clusters" menu at the top of the screen and selecting "Swan Shell Access"
- Change to your $WORK directory and run "git clone https://github.com/unlhcc/hcc-ai-ml-workshop-2025.git" without the quotes. Verify that the data was downloaded using the "ls" command.
- Return the the OOD main page and start a Jupyter Lab session from the "Interactive Apps" menu using the settings in the section below.
- Navigate to the "hcc-ai-ml-workshop-2025" directory on the left hand side of the screen.
- Open the "prepare_datasets.ipynb" notebook and run all cells
- When all cells have finished execution without error, put up your YELLOW sticky note.
Open OnDemand Settings:
| Parameter | Value |
|---|---|
| Jupyter Lab version | 4.0 |
| Working Directory | /work/<group name>/<user name> |
| Number of cores | 8 |
| Running time in hours | 3 |
| Requested RAM in GBs | 32 |
| Partition selection | gpu |
| Reservation | Located on the back of your name tag |
| GRES | gpu |
| Job Constraints | ---Leave BLANK--- |
Resource Reservation for GPUs: Located on back of nametags
JupyterLab Settings:
| Parameter | Value |
|---|---|
| INSIDE JUPYTER LAB: Kernel Name | hcc-ai-ml-workshop-2025 |
Useful links and commands for the workshop:
Workshop Notebooks Repository
Intro to ML and AI Slides
ML/AI Workflows
Introduction to PyTorch
\/\/\/\/\/\/\/\/FEEDBACK\/\/\/\/\/\/\/
Workshop Feedback
/\/\/\/\/\/\/\/FEEDBACK\/\/\/\/\/\/\/\
Bash Cheatsheet (External)
HCC info sheet
HCC SLURM scheduler cheat sheet
Please contact us at hcc-support@unl.edu if you have questions or concerns and we will be happy to help!