1
00:00:00,580 --> 00:00:06,970
So welcome to a brand new module this module is going to be pretty jam packed with some pretty cool

2
00:00:06,970 --> 00:00:08,110
stuff as well.

3
00:00:08,140 --> 00:00:11,050
So what's coming up in the previous module.

4
00:00:11,050 --> 00:00:13,190
We introduced the multilayer perception.

5
00:00:14,020 --> 00:00:20,620
And it didn't really do all that well on the classification task but it was given.

6
00:00:20,620 --> 00:00:24,070
So this time round we're gonna give it another chance.

7
00:00:24,250 --> 00:00:25,360
We're gonna see how it does.

8
00:00:25,360 --> 00:00:28,840
On a different dataset what data set is that.

9
00:00:28,840 --> 00:00:35,070
Well it's gonna be a data set of handwritten digits the so-called enhanced dataset.

10
00:00:35,290 --> 00:00:40,410
My motivation for continuing on with the perception is that we get to understand it on a deeper level.

11
00:00:40,630 --> 00:00:44,960
And we also can learn some proper tensor flow along the way.

12
00:00:45,160 --> 00:00:51,100
And when I say proper tensor flow I mean taking off the caris training wheels and looking under the

13
00:00:51,100 --> 00:00:57,520
hood we're gonna be getting our hands dirty with the tensor flow graph tensor flow sessions forward

14
00:00:57,520 --> 00:00:59,020
tensor baud goodness.

15
00:00:59,110 --> 00:01:05,410
And yes you guessed it tenses for this module are going to be running everything locally on my machine

16
00:01:05,890 --> 00:01:12,130
but you can also follow along in the Google collab notebook the architecture that I'm going to be proposing

17
00:01:12,130 --> 00:01:19,240
for our neural network is not going to require a GP you it's not gonna be too demanding but depending

18
00:01:19,240 --> 00:01:25,060
on your machine and if you do need more juice or you want to experiment with a more demanding version

19
00:01:25,060 --> 00:01:31,750
of the perception then Google cool up notebooks and their free GP runtime is gonna be your friend.

20
00:01:32,110 --> 00:01:37,510
So if you need the juice go with the Google cool lab notebook and if you'd like to see what's going

21
00:01:37,510 --> 00:01:42,970
on with tensor board and we're going to be working a whole lot with tensor board in this module then

22
00:01:43,210 --> 00:01:45,310
run everything locally.

23
00:01:45,350 --> 00:01:49,480
Now without further ado let's get set up and let's get started.