Running Python Data Science Docker Images in AWS with NVIDIA NGC

Показать описание

NVIDIA NGC allows you to run premade Docker containers for many common data science tasks. This allows you to leverage transfer learning and have all the drivers already preconfigured. In this video, I show how to launch NGC containers inside of AWS. NGC is available entirely within the Amazon AWS marketplace, with no markup. As an example, I launch the RAPIDS container.

2:17 NGC on Local Instance
3:00 NGC on AWS Marketplace
4:22 Launching a Spot Instance for NGC
6:14 EC2 Login Key Pair
6:55 Logging in, Putty and Tunneling
7:44 Connecting to your NGC EC2 Instance
8:14 IAM and AWS Configure
10:48 Loading the NGC Container
14:26 Launching Jupyter Lab

NVIDIA NGC

NVIDIA NGC AWS Marketplace

Follow Jeff Heaton/Subscribe:

Рекомендации по теме

Комментарии

Jeff! You saved me a lot of efforts. I was going to make a video on this, now I'll just add it in my playlist. Thank you!
- Sanyam

ChaiTimeDataScience

Love the practical vids, thumbs up to future vid on using s3

zackstrater

Definetly interesting since it saves time and a lot of local maintenance all the time when updates are released. Thumbs up!

fredrikhansen

That's amazing to hear. Please make video about how to use CUDANN in local computer using docker image.

MahirAshabAhmed

Jeff - Great videos. My team out in SF loves the channel. Quick question - When navigating to the NVIDIA NGC Catalog and selecting the RAPIDS toolset, the page says "This version has been removed and is no longer available to new customers" - is RAPIDS no longer supported on AWS? I also noticed on the BlazingSQL site that BlazingSQL has been deprecated with release 21.12 in favor of Dask-SQL. What does this mean for the future of BlazingSQL?

smm

Great video! Are you thinking on doing a video of SageMaker?

brunomanuelcavagnaroolcese

Will Nvidia Omniverse have anything like artbreeder, with the ability to blend, and cross breed? Thanks.

scotthaney

Thanks, Can I do same with deep stream on aws

michaelosumune

That doesn't make any sense Jeff, its easier to put the parts together.

RevivedWarrior

nice video. I assume In this case we can bypass ECR if the RAPIDS image is downloadable directly from nvcr.io/nvidia/rapidsai My guess is that an AWS image has a better integration with IAM but in our case we still had to define credentials to authenticate against ECR

alex-gqm

Running Python Data Science Docker Images in AWS with NVIDIA NGC

Running Python Data Science Docker Images in AWS with NVIDIA NGC

Docker for Data Science tutorial | Docker for Python

Matthijs Brouns - 10x smaller docker containers for Data Science | PyData Eindhoven 2020

Docker + Python for Data Science and Machine Learning | Real Python Podcast #8

Jupyter Notebook using Docker for Data Science (Demo)

Complete Dockers For Data Science Tutorial In One Shot

Containerize Python Applications with Docker

How to “Dockerize” Your Python Applications | How To Build And Run A Python App In Docker Container...

Running ChatGPT Locally: A Guide to Open WebUI and Ollama Integration

Docker For Data Scientists

Data Science Workflows using Docker Containers | Future of Data & AI | Data Science Dojo

Docker Tutorial 6- Writing, Building And Running Docker Image| Machine Learning

Docker for Data Science: Deploying a Web Application

How to run Python scripts in a Docker container

Trying Docker for containers with Python and Google TensorFlow machine learning data science

Docker for Data Science - Aly Sivji

Docker Data Science Pipeline

Docker Simply Explained with a Machine Learning Project for Beginners

How to Convert your Machine learning Model into a Docker Image (Python ML Model)

'Docker, Python, and You' by Matt Rasband

Data Science Workflows using Docker Containers

Docker data science pipeline

2. Dev setup — Docker crash course for data scientists #CH4DS

100+ Docker Concepts you Need to Know