Registration for the Machine Learning Bootcamp is now closed.

Thank you for the immense interest in the bootcamp. We have received a large number of requests, and will be reaching out to the first 50 respondents. We will provide information about any future trainings when available. If you have any questions or need more information, please contact us.

Original Notification

On Wednesday, May 20, 2020, NAS will host a one-day Machine Learning Bootcamp featuring an introduction to GPU computing and RAPIDS, followed by introductions to cuDF and cuML. We will also get started on Dask and Deep Learning on GPUs. There will be several opportunities for lab exercises and questions and answers. See the detailed agenda below.

The event will be held online/remote via a WebEx meeting.

To register, please send your name and affiliation to support@nas.nasa.gov with "Machine Learning Bootcamp" as the subject.

Okta's Advanced Server Access client will be required on your system to connect to the training cluster.

NASA ML with RAPIDS Bootcamp Agenda

Note: All times are Pacific Time.

Welcome and Cluster Setup (9:00 – 9:15 a.m.)

Introduction to GPU Computing and RAPIDS (9:15 – 10:00 a.m.)

     Machine Learning with RAPIDS

• Avoid Data Movement Once on GPUs
• Evolution from Hadoop/Spark to RAPIDS
• How to Access the Software and Get Started

Introduction to cuDF (10:00 – 10:45 a.m.)

• Overview
• Scaling with Dask for Large Datasets
• CuPy, Numba, and PyTorch Compatibility
• GPUDirect Storage
• Hands-on:
  • Pandas vs. cuDF (GPU-accelerated dataframe manipulation)

Break (10:45 – 11:00 a.m.)

Introduction to cuML (11:00 a.m. – 12:00 p.m.)

• Overview
• API and Available Algorithms
• A Look at the Python and C++ Implements
• Benchmarks
• Hands-on:
  • Scikit-Learn vs. cuML (GPU-accelerated machine learning)
  • Utilize Machine Learning Algorithms, such as Linear Regression and UMAP

Break (12:00 – 1:00 p.m.)

Introduction to Dask + GPUs (1:00 – 1:30 p.m.)

• Setup Dask Client and Workers, and Execute a Parallel Program

Introduction to Deep Learning on GPUs (1:30 -2:00 p.m.)

• Introduction to CuPy and Numba
• Overview of Deep Learning Institute (DLI): Online Courses and In-Person Workshops
• Open Q&A Session