Overview
Watch a 20-minute tutorial demonstrating how to achieve a 25x performance boost in XGBoost training, inference, and feature engineering using RAPIDS on NVIDIA GPUs within Google Colab. Follow along with real-world taxi fare data to build baseline models and enhance model accuracy through advanced feature engineering techniques. Learn to enable GPU acceleration, construct baseline and XGBoost models, and compare performance with RAPIDS enabled and disabled. Access pre-installed RAPIDS and XGBoost environments directly in Google Colab, with a provided notebook for hands-on practice. Master GPU-accelerated machine learning workflows while exploring practical applications in data science and model optimization.
Syllabus
00:00 - Introduction
0:50 – Enable GPU in Google Colab
6:10 – Build baseline model
10:05 – Build XGBoost model
16:11 – RAPIDS On/Off
19:46 – Conclusion
Taught by
NVIDIA Developer