Gpu Based Docker Containers With Python_Part-1

Building a Docker image is generally considered trivial compared to developing other components of a machine learning system like data pipeline, model training, serving infra, etc. But an inefficient, bulky docker image can greatly reduce performance and can even bring down the serving infra.

In this video series, we would go over a hands-on experience with Erdem

If you are new to docker or if you want to work on GPU such as projects like
TensorFlow, This is one of the good videos that can work on a Python-based library
then this video might be the correct place for you to install a Python Library into
a Docker file. Our goal is a bit complex [Many small parts getting together to run
a system]. But as far as I see if you look at the Teslabot or the poker ai, You are going to see that the AI-based workloads are coming to the cloud and coming to more and more use cases and we are going to see these use cases much more in the future.

Topics Covered :
0:01 Intro
0:54 How we can build a file using Python & GPU based Resources
1:19 How you really can get a Hands-on DevOps experience, Creating a Linux container
1:48 Focus on Nvidia Cuda Cores
2:01 Using an Nvidia GPU
2:08 Steps taken in the process
2:19 Using Pomodoro techniques
2:31 What is a Pomodoro?
2:33 Pomodoro technique 25 minutes timed out
2:38 Objectives and Key results
3:05 Different docker images in action
3:35 Pros and cons
3:39 Generic images
4:03 First Principles approach
4:10 Premise and Conclusion
4:23 Plain images from Nvidia or Custom Image Provider
4:42 Examining the Docker Files as many as you can
5:07 Documentation is part of the job
5:33 Ideas and Proof of concepts
5:53 Concepts drivers container runtime
6:00 Dockerhub.com
6:32 github.com
6:48 Confirmations are Vital try to be less wrong
7:05 Azure support's great help in the process
7:21 Stay active on these Platforms: Reddit and Github
7:23 Nvidia, Python, Linux, Docker coming together in the Process
8:16 Docker is a great solution to get rid of Snowflakes
8:30 Infrastructure as Code [IAC]
8:45 Nvidia based container images
8:53 Cloud vendors [Azure, AWS, GCP] come to the rescue in GPU Shortage
9:28 Common Libraries: such as software properties
9:42 Repos dependencies Python Pip
10:31 Docker builds an operating system mostly to run a process
11:16 Pip package libraries
12:31 Cached layers in Docker

Subscribe and Ask questions for the future Videos

GET MENTORING FROM ERDEM
https://superpeer.com/rifaterdemsahin


MY FREE ONLINE COURSES:
? Etoro Traders Tips and Tricks- https://www.youtube.com/watch?v=bHvJ-Zv8etE&list=PLEaC7OEmKSrfZZqdwbu2SxXAil0SBtix2

MAKE MONEY

? Active Portfolio on Etoro to copy strategy https://www.etoro.com/people/rifaterdemsahin

BE MY FRIEND:
? My website/blog - https://www.devops.engineering
? Twitter https://twitter.com/rifaterdemsahin
? Linkedin Profile - https://www.linkedin.com/in/rifaterdemsahin
? Facebook https://www.facebook.com/groups/394617277679731/

WHO AM I:

? Starter ? Geek ? Curious ?Talkative
Ultimate Contractor!

I'm Erdem, a DevOps contractor living in Cambridge, UK. I create content on technology, finance, psychology and lifestyle design. I am an active learner and love to share new ways to learn things and believe in self-learning. I create courses for people to escape poverty by showing them how to close the skills gap.