Python Essentials for Scientific Computing/High-Performance Computing

Free one-day Python hands-on workshop covering topics related to leveraging high-performance computing architectures for machine learning. Learn common techniques for extracting meaning from data using matplotlib, seaborn or plotly in the data visualization session. Explore how to turn your Python codes into distributable packages in the Python packaging session. Learn to use the Dask library for parallel and distributed computing for easy scaling of common Python packages like NumPy, pandas and scikit-learn. And learn basics of JAX library with a focus on jax.numpy API and how JAX can use accelerators. Attendees will get hands-on experience using UCF’s Advanced Research Computing Center (ARCC) HPC resources during the workshop. This is a full-day event with five sessions and lunch will be provided. The workshop is open to all UCF graduate students, post-docs and faculty. This is an in-person event.

Speakers: Benjamin Keene and David Wright

Audience level - Intermediate:

• Should have a basic familiarity of the Linux command line
• Should be moderately skilled in Python and familiar with Pytorch
• Familiarity with the field of machine learning would be helpful

Workshop Program

9–9:30 a.m. (30 min): Introduction and ARCC setup for hands-on workshops
9:30–10 a.m. (30 min): Showcasing scientific usecases and applications
10–11 a.m. (1 hr): DASK hands-on 
11–11:15 a.m. (15 min): break
11:15–12 p.m. (45 min): ML workflow optimization with pytorch (hands-on) Part 1
12–12:45 p.m. (45 min): Lunch will be provided
12:45–1:30 p.m. (45 min): ML workflow optimization with pytorch (hands-on) Part 2
1:30–3 p.m. (1.5 hr): Data visualization with Matplotlib and Seaborn
3–3:15 p.m. (15 min): break
3:15-3:45 p.mm (30 min): JAX hands-on
3:45–4:30 p.m. (45 min): Python packaging tools and practice

Temporary accounts on UCF’s ARCC machines will be provided and workshop materials will be distributed to attendees. Event will not be recorded.

Parking information: Parking is available in the Digital Learning Center lot, with overflow at University Tower if needed. No backing into spaces is allowed. If you do not have staff parking, the charge will be $5 for the day.