You’ll be calling of the same point to get your predicted models Right So I’m gonna call this and is going to give you predictions back That’s how exactly the entire life cycle fronts enough Okay so that’s how it works on just going to stop pushing this because we may not have enough time to cover the country Be maybe All right so that being said you go statically jump in tow over Pagemaker positions and things like I just created some backup environment as well In case you guys know how exactly the musicians would be right It may not work exactly the time when you really need it Just read it.
Some backups Ah environment somewhere so that we are all good Um first and foremost what I want to show you guys is um is there an area or a region where I can start everything together Okay let’s start with okay Sagemaker services Very well I get it I’m assuming everyone knows how to get into our this loveliest um um gun soul providing you Zaidi parcelled and get into AWS console You guys are an expert on the and premature I selected by not California region Right The reason behind nothing specific I just want to have my demo started over here but I’m going to switch back to an environment where I already created and kept it something so that we don’t waste time doing that.
I’m in my north California now on da Amazon Page maker I’m going to happen Amazon Um sorry I missed if it’s better Just just put in stage Maker Right-click on stage makeup which is there on my left inside off my screen All right It’s going to say Amazon Sage maker Bill train and deploy machine learning models that scale Right That’s exactly what you can see There’s a bill processes trained process There’s a tuning Piper Pereira tries parameter tuning That’s also even discussed on that I can deploy my model Right Make sure you’re realizing that image with us spoken about um sometime back All right So where do I start my journey with Click on Notebook Instance Now what exactly is This is the place Very ago create your Jupiter notebook All right So how do I do that Click on create notebook Instance.
I’m going to see my notebook Instance Name great learning notebook I just want to have not can um California N c o year All right Notebook is coming back and asking You Okay How big a machine you need right Does your re a data-crunching or do you have a humongous amount of data with you Your data engineering process is going to take you Does that make you go for a massive bigger machine Are Are you okay with a smaller machine That’s a good question It’s going to ask you So that’s where you can see over here Understand One thing folks This is not a machine that is responsible for training right Let me just come back to that slide once again So they spark Not this part for training.
There is a different machine altogether going to come up So first to write your code What is the machine that you required Now you may have your Christian So why do I need to have a massive big machine just for writing my court Understand Even your data engineering your data of the future engineering whatever manipulation whatever the data massaging their crunching all those things what you want to do that should be done in this mission That’s the reason you have to choose an appropriate machine which can actually do that Us for you I’m going to pick up them two medium Um I would request you guys to just go for Arab list on Amazon dot com slash free with say’s for two months That is one machine available for you I am assuming that’s going to be them to medium We don’t have any to micro missions over a year Right Um I would assume it should bet two medium so that you also can exactly do this exercise by yourself without I mean free Of course.
That’s exactly what I mean Um now the permission Bart off it So you may have to create an Amazon Page maker role in case if you don’t have one you’re just going to say create a new role option and you’re going to select Rich s three-bucket You’re interested already can create all s three-bucket I do already have a role in place so I don’t need to worry about it And my assumption is you guys are pretty much a veil of what exactly the role and things like that If you have in this three-bucket where your data is reciting you can even specify only that bucket name over here are I’m going to say any estate market that means every single Stephen could you have access Troy I don’t want to do that right now.
I do have Aziz rules already available Okay um I’m gonna stick on to complete default mode over here on What I’m going to do is create a notebook instance Click on this That means at T two medium machines is going to come up for me Okay On this process is going to take some time for you Um is going to take some time before that’s our first process That’s got completed building up selected the medium What is a meter on what basis I can decide whether to select medium or large or extra-large something Yeah so that visibility you should be having basically in cases here If your data is going to be pretty heavy and heavy or huge on then you’re going to see my data crunching for doing my data crunching I may need a bigger machine Right So you know that there are t two medium d 28 x large for X slash sort of machines.
I need to understand for turning my data one GB ram on one core processor That’s not enough right So that kind of inside you should be having yourself because otherwise what’s going to happen It’s going to take You have a massive huge amount of time doing it on even at times What happens if your entire data needs to decide in memory even it can fail So so that decision you should be knowing based only a cloud expert eyes So that’s where you need to So just go and see the configuration 42 medium and see How much is if it’s going to be to score four-core processor and TV but now when you’re going to have around Finder gene AUF data to be played around in this environment is going to say.
Furthermore, I’m not happy with you night So that’s exactly where you are going to make a decision No I cannot survive in D two medium machines with 500 GB of data I may want to go for a bigger machine so that I can get my data engineering process a region in the training process being processed much faster otherwise going to take you almost 45 hours for your data crunching crosses Okay, but this is only during development Absolutely Yes This is your development time Great Any Christians or can we move forward No instance Right.
It says you’re Mel dirty too So normally we don’t see this Remember Is this specific to mention numbing That’s Car rick that you will be allowed to pick up only those machine led So this is customized for machine learning So behind architecture is made in such a way that your machine learning kind of an engineering process Ah these machines are customized for that So we would not recommend you to pick up a T two medium or ah, ah I like a lot of that C four or C five machines run this pick up machines only with start with ml, not Darla That starts a kind of mission I think there is an awful for reportedly what is that for on Is there in your for Yes you had okay let me just real quick you that part so you can get your code bringing up from getting repository aspirin.