Explore high-speed data processing using NVIDIA RAPIDS on AWS Cloud in this 21-minute video tutorial. Learn how to leverage GPU-accelerated Pandas-like dataframes for fast preprocessing and seamless integration with GPU-based models. Follow step-by-step instructions to set up an AWS environment, including S3 bucket creation, IAM access configuration, EC2 instance setup with spot pricing, and EBS volume management. Discover how to access NVIDIA NGC containers, use Jupyter notebooks, and implement RAPIDS for efficient data manipulation. Gain insights into utilizing XGBoost on NVIDIA GPUs for machine learning tasks. Perfect for data scientists and engineers looking to optimize their workflow with cloud-based GPU computing.
Overview
Syllabus
AWS File Systems
RAPIDS on Real Data
Kaggle Otto Group Challenge
Create an AWS S3 Bucket
Sending Data to AWS S3
Setting up AWS IAM Access
Create an AWS IAM Role for AWS EC2
Create an AWS IAM Policy
Create a new AWS EC2 Instance with spot
Connect to the Running EC2 Instance with Putty and Tunneling
Starting Docker on AWS EC2
Creating an AWS EBS Volume for EC2
Mount and format EBS Volume
Access NGC Images
Jeff's AWS Setup Chart
Accessing Jupyter NGC Instance
Using RAPIDS
Using XGBoost on an NVIDIA GPU
Taught by
Jeff Heaton
