WEBVTT

00:00:00.110 --> 00:00:02.910
>> We finally have
our environment setup

00:00:02.910 --> 00:00:05.220
both in the Cloud and locally.

00:00:05.220 --> 00:00:08.895
We have our machine-learning
resource up on Azure and

00:00:08.895 --> 00:00:11.040
Visual Studio Code now has all

00:00:11.040 --> 00:00:13.365
of the packages and
extensions we need,

00:00:13.365 --> 00:00:17.385
like Python, Jupyter Notebooks
and Azure Machine Learning.

00:00:17.385 --> 00:00:21.150
Next step is to make sure that
our local environment is actually

00:00:21.150 --> 00:00:24.900
connected to our unique
machine-learning resource in the Cloud.

00:00:24.900 --> 00:00:27.000
As always, don't forget that we have

00:00:27.000 --> 00:00:28.980
additional resources linked on

00:00:28.980 --> 00:00:32.340
the screen in front of you or
in the description down below.

00:00:32.340 --> 00:00:34.950
First, I'll head into
the machine learning

00:00:34.950 --> 00:00:36.960
resource on Azure and download

00:00:36.960 --> 00:00:39.530
the config file that gets
automatically generated and

00:00:39.530 --> 00:00:42.325
save it to the same folder that
I have my Jupyter Notebook in.

00:00:42.325 --> 00:00:43.845
Now, when I open VS code,

00:00:43.845 --> 00:00:45.405
I can see that config file.

00:00:45.405 --> 00:00:48.290
It's a very simple one that
just has my subscription id,

00:00:48.290 --> 00:00:50.770
resource group, and workspace name.

00:00:50.770 --> 00:00:53.480
From Azureml-core, I need to

00:00:53.480 --> 00:00:56.135
import the workspace
and experiment classes.

00:00:56.135 --> 00:00:58.190
Now, I'm going to create
a local variable that'll

00:00:58.190 --> 00:01:00.335
point to my workspace up on Azure.

00:01:00.335 --> 00:01:03.240
I'll use the config file for this.

00:01:04.280 --> 00:01:07.025
Next, I'm going to give my experiment

00:01:07.025 --> 00:01:11.880
a name and next I'm going
to create a new experiment.

00:01:11.990 --> 00:01:14.000
Notice that you might get

00:01:14.000 --> 00:01:17.540
another request to
authenticate with Azure.

00:01:17.970 --> 00:01:21.470
If you do, go ahead and do so.

00:01:21.630 --> 00:01:25.925
Next, I'm going to
create my CPU cluster.

00:01:25.925 --> 00:01:28.460
I want this CPU cluster to be

00:01:28.460 --> 00:01:29.690
an Azure Machine Learning

00:01:29.690 --> 00:01:32.900
compute and I need to have
access to the compute target.

00:01:32.900 --> 00:01:36.540
Let's import both of
those classes first,

00:01:37.520 --> 00:01:43.120
I'm going to give my CPU cluster
a name. I'll keep it simple.

00:01:43.610 --> 00:01:48.065
I'm going to create a
configuration for my CPU clusters.

00:01:48.065 --> 00:01:50.330
Notably, I want to
specify the VM size.

00:01:50.330 --> 00:01:53.430
I'm just going to choose a
standard size for this one.

00:01:53.450 --> 00:01:58.360
Then I'm going to create
my compute target.

00:01:58.640 --> 00:02:01.700
Finally, I'm actually
going to send this up to

00:02:01.700 --> 00:02:04.010
my workspace in Azure and

00:02:04.010 --> 00:02:06.260
specifically I want
this cell to continue

00:02:06.260 --> 00:02:09.170
executing until it's
been completed in Azure.

00:02:09.170 --> 00:02:12.930
I'm going to use the wait
for completion function.

00:02:14.840 --> 00:02:18.760
Let's create our CPU cluster.

00:02:20.510 --> 00:02:24.590
A couple of neat things about
this is once it's been completed,

00:02:24.590 --> 00:02:26.525
we can see it in two
different places.

00:02:26.525 --> 00:02:30.010
One is in the Azure extension
right here in Visual Studio Code.

00:02:30.010 --> 00:02:33.370
If you click on it and open up
your pay-as-you-go subscription,

00:02:33.370 --> 00:02:35.950
you should find your
machine learning resource

00:02:35.950 --> 00:02:39.110
and in there you should find the
compute that we just created.

00:02:39.110 --> 00:02:41.910
Here's my CPU cluster.

00:02:42.050 --> 00:02:44.780
You can also find this on

00:02:44.780 --> 00:02:48.530
Azure by going to the
machine-learning resource

00:02:48.530 --> 00:02:51.940
and clicking ''Compute'' you'll

00:02:51.940 --> 00:02:54.200
be able to find this
under training clusters.

00:02:54.200 --> 00:02:56.060
Since this is a clusters specifically

00:02:56.060 --> 00:02:58.880
used for Azure Machine
Learning training.

00:02:58.880 --> 00:03:02.100
I think we're ready for the data now.