NVIDIA Holoscan Bootcamp

Authors: Denis Leshchev (NVIDIA), Adam Thompson (NVIDIA), Nicolas Lebovitz (NVIDIA)
Supported platforms: x86_64, aarch64
Last modified: March 18, 2025
Language: Python
Latest version: 0.1.0
Minimum Holoscan SDK version: 2.2.0
Tested Holoscan SDK versions: 2.3.0
Contribution metric: Level 1 - Highly Reliable

Overview

This is a meta-tutorial for deploying and running NVIDIA Holoscan Bootcamp lab materials.

Description

NVIDIA Holoscan Bootcamp was conducted by the NVIDIA team in collaboration with OpenHackathons on September 24th, 2024 (event link). The bootcamp included a lecture and a hands-on Python-based lab. The lab materials are freely available for downloading at OpenHackathons repo.

NVIDIA Holoscan Bootcamp lab materials include a Jupyter notebook for self-paced studying of the core Holoscan concepts.

The notebook covers the following topics: - Creating custom Holoscan operators - Connecting operators to form an application - Build Holoscan applications with GPU-accelerated Python packages - How to build a sensor to visualization pipeline - Measuring and optimizing application performance - How to build complex sensor processing workflows using multiple fragments and the advanced schedulers

The notebook guides users through the topics with example code and diagrams, and includes exercises to help them practice the concepts they've learned.

To run the lab materials, clone the repo and follow the deployment guide.