Serverless Computing with Google Cloud

preview_player
Показать описание
Serverless computing is a cloud-computing model where the cloud provider acts as the server, dynamically managing the allocation of machine resources. What that means is 'pay-as-you-use' pricing for the developer for their computing tasks. In this video i'll walk through some useful parts of the Google Cloud suite of services and use it to help train+test a wide+deep neural network to predict customer purchasing power. Topics covered include cloud computing models, kubernetes, Colab, and regression models. Enjoy!

Code for this video:

Please Subscribe! And like. And comment. That's what keeps me going.

Want more education? Connect with me here:

This video is apart of my Machine Learning Journey course:

More Learning Resources:

Join us in the Wizards Slack channel:

Sign up for the next course at The School of AI:

And please support me on Patreon:
Signup for my newsletter for exciting updates in the field of AI:
Комментарии
Автор

I'm learning python as a hobby at my very own snail pace. Really enjoy your stuff.

VitePapa
Автор

Thanks for bringing out the content that will power our future

GabrielScindian
Автор

Great overview of GCP with the focus on ML

web
Автор

Thanks for sharing. Excellent job on this video. I just wanted to let you know that I am grateful for you. You got incredible heart and one of the best I know in this field.

zillionairetc
Автор

I've been considering moving to GCP, and boom Siraj uploads a video about it. Thanks Siraj.

tjnaz
Автор

Nice but would like to see a long form version of this video - these are great for starters but your left hungry after.

ravisawhney
Автор

Please!!!
Make video on how to use Google colab!!

rushikeshdarge
Автор

Thanks for this Siraj. You always rock :)
Don't you need to include FaaS (like GCP functions, AWS lambda etc., ) ?

shashankbarki
Автор

Well, yes, okay serverless are a neat little thing, however they DO NOT solve all the mentionned problems.

If you develop any app, you are still responsible to update it and patch it.
- Libraries that you use might have security holes discovered or undiscovered yet.
- Crypto will evolve (ALWAYS), as computer get more and more power and techniques become more sofisticated crypto adapts often. It's your responsibility or your library responsibility to adapt (ence you, because you control the version of your dependencies).
- You still need to manage monitoring, monitoring is NOT an option specially on serverless if you don't want to end with an insane bill.
- Careful with the solutions you choose, you might be locked in to a specific solution quite easily.

If you want to go serverless or app engine style and don't want to be locked in :
- Select any Kubernetes service (or deploy your own if you know what you are doing)
- Use any FOSS serverless solution that has kubernetes and containers in mind

A few solutions for serverless on top of kubernetes :

Again, you are never free of maintaining your app, that doesn't exist. Even on the dumbest wordpress, you might end up with addons that you need to update yourself and js libraries that you need to update to avoid being on the front page of hacker news.

YvanDaSilva
Автор

Don't do... Hit and run videos... Do proper series.... On... Any single topic... Learn from coding train

darogajee
Автор

Can you please cover how to train and run R model in ML engine. Also if I have to use app engine flex to do the same what are the best practices

codersgarage
Автор

Equivalent Platform for the cool kids doing PyTorch ?

TheFakeFirang
Автор

Identifying clouds in satellite images.

cloudiness = |image - regression of image to typical perlin noise|?

Just an idea.

alanhere
Автор

Siraj, is it sponsored video? what about decentralized apps options?

islepcov
Автор

Can you please make a video about designing graphics throught AI like book cover

TheSukhdeep
Автор

#notifications_squad
thanks siraj !!😊

meetshah
Автор

Google Collab free? Nothing is free. The trade is something more valuable than money: ideas/code. But then again, we do that all the time with even this comment section on YouTube and every other "free" platform. The question is, is your innovation worth the trade for technological use. If it is, then it's a great option. 👍

DragusCalliforous
Автор

Thanks! Does serverless computing only allow you to use Google's service? Like Google NLP, ML things like that? OR I can also run my own model from scratch?

minjun
Автор

we need a serious comparison between all cloud services, don't we?

ibrahimhalouane
Автор

Is it possible to offload some compute from my laptop to a remote server to speed up my system? Like, I'm playing some graphics heavy game on steam or blizzard, and I have good internet speed, could I use someone else's GPU that's somewhere else on the planet, or maybe utilize a remote GPU cluster. How can I do that?

NickKartha