TPC26 Tapis Tutorial
Tutorial Materials
Tutorial Slides
Slides for this tutorial
📝 Reservation Info
Allocation Code:
Sunday sessions:
Monday sessions:
Slides for this tutorial
📝 Reservation Info
Allocation Code:
TRA24006Sunday sessions:
Tapis+Tutorial-SunMonday sessions:
Tapis+Tutorial-Mon
Trillion Parameter Consortium 2026 Tutorial Schedule
Day 1 (May 31st) Schedule
| Time | Duration | Description | Speaker |
|---|---|---|---|
| 2:00-3:00 PM | 60 min | (Lecture) Section 1: AI/ML, the National CyberInfrastructure Ecosystem and Tapis | Anagha Jamthe |
| 3:00-3:30 PM | 30 min | (Hands-on) Section 2: Initial Steps With TapisUI | Christian R. Garcia |
| 3:30-4:00 PM | 30 min | Coffee Break | |
| 4:00-4:45 PM | 45 min | (Lecture) Section 3: Modern AI, Scientific Workflow and Infrastructure Gaps, Tapis Solutions | Wei Zhang |
| 4:45-5:30 PM | 45 min | (Hands-on) Section 4: Running AI Inference on Vista Via Tapis and FlexServ | Wei Zhang |
Day 2 (June 1st) Schedule
| Time | Duration | Description | Speaker |
|---|---|---|---|
| 9:00-9:30 AM | 30 min | (Lecture) Section 5: Computer Vision Models in Scientific Applications | Christian R. Garcia |
| 9:30-10:00 AM | 30 min | (Lecture + Hands-on) Section 6: Fine-tuning the Vision Model with Ultralytics and Tapis | Anagha Jamthe |
| 10:00-10:15 AM | 15 min | (Hands-on) Section 7: Launching Jupyter Notebook | Christian R. Garcia |
| 10:15-10:30 AM | 15 min | (Lecture) Section 8: Prompt-Engineering: Best Practices | Anagha Jamthe |
| 10:30-11:00 AM | 30 min | Coffee Break | |
| 11:00-11:30 AM | 30 min | (Hands-on) Section 9: Generating Code for a Vision Task and Executing in Jupyter | Wei Zhang |
| 11:30-12:00 PM | 30 min | (Hands-on) Section 10: Evaluating fine-tuned model within Jupyter | Wei Zhang |
| 12:00-12:30 PM | 30 min | Additional Use Cases and Discussion | Christian R. Garcia |
The resources you will be using today are provided by funding from the National Science Foundation.
Intro to Tapis
Tapis Frequent Questions
What is Tapis?
Tapis is an NSF-funded API platform for managing high-performance computing (HPC) systems, running jobs, and managing data across HPC centers. Tapis allows users to manage compute resources, submit jobs, manage files through APIs or TapisUI.What is Flexserv?
Flexserv is a Tapis service that allows you to deploy and serve AI/ML models on HPC or VM resources. It provides inference endpoints and a graphical interface for managing the service.Where can I ask questions after the tutorial?
- During the tutorial: Raise your hand and a taccster will come help you.
- After the tutorial: Open a ticket at https://portal.tacc.utexas.edu/tacc-consulting or post in the Tapis Slack community.
- GitHub issues: For bugs or feature requests, visit https://github.com/tapis-project.
