File size: 4,442 Bytes
d157f08 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 | 1
00:00:01,580 --> 00:00:07,310
And welcome to chapter 8 point 1 0 where we're actually going to use curus to display a visual output
2
00:00:07,460 --> 00:00:08,420
or model.
3
00:00:08,650 --> 00:00:12,950
So remember before previously I was drawing all this nice visualization of our model.
4
00:00:13,220 --> 00:00:18,640
Well carrots can actually do something not quite as nice as this but it produces a pretty decent model
5
00:00:18,660 --> 00:00:22,730
visualization that helps you basically show people and explain your model.
6
00:00:22,910 --> 00:00:24,240
So let's see how we do it.
7
00:00:24,290 --> 00:00:26,220
Let's go back to what I thought in the book.
8
00:00:26,500 --> 00:00:26,840
OK.
9
00:00:26,870 --> 00:00:32,630
So now we're about to visualize our model so to visualize our model we need to import this library from
10
00:00:32,680 --> 00:00:34,200
Carousel dysfunction.
11
00:00:34,280 --> 00:00:39,680
It's called plot model and it's found in Cara's utilities thought visualization utilities vid's underscore
12
00:00:39,680 --> 00:00:41,220
utilities for short.
13
00:00:41,240 --> 00:00:43,980
So what we do we create or recreate or model first.
14
00:00:44,030 --> 00:00:49,430
We don't necessarily have to do this but it's good practice just in case we didn't do it before and
15
00:00:49,450 --> 00:00:53,370
it previously as in the previous cells and this I Pitre notebook.
16
00:00:53,390 --> 00:00:59,800
So let's go ahead and run the slime on this block and we get this table which is our same model output
17
00:00:59,810 --> 00:01:00,610
from before.
18
00:01:00,770 --> 00:01:01,740
All right.
19
00:01:01,970 --> 00:01:04,370
This in itself is a pretty decent visualization.
20
00:01:04,370 --> 00:01:06,970
However it's not like a visual diagram.
21
00:01:07,220 --> 00:01:09,380
What we're going to do is produce a visual Vaga now.
22
00:01:09,440 --> 00:01:15,320
So to do this we use a plot model function and a plot modeled function basically takes a model that
23
00:01:15,340 --> 00:01:20,820
we find here we take we enter a pot for the file to be saved.
24
00:01:20,870 --> 00:01:25,550
So we just use this path here which is the way our train models are saved and we give it a phylum model
25
00:01:25,550 --> 00:01:30,720
and scale plot PNB we can be more descriptive and give it like this model.
26
00:01:30,740 --> 00:01:31,510
All right.
27
00:01:32,060 --> 00:01:34,860
And then after we use matplotlib to actually show.
28
00:01:34,880 --> 00:01:40,220
So we just point matplotlib here that image directory of where we see that and we plotted here and it
29
00:01:40,220 --> 00:01:42,140
comes up below right here.
30
00:01:42,200 --> 00:01:43,500
So let's see how this works.
31
00:01:43,520 --> 00:01:48,990
Let's run this block and then we go here's a nice model visualization.
32
00:01:49,040 --> 00:01:54,530
What's cool about this is that we have inputs and outputs coming in and out is actually quite nice.
33
00:01:54,530 --> 00:01:55,970
This is basically a random number.
34
00:01:55,970 --> 00:02:00,990
Never have been able to figure out what this number actually is pretty much ignored for now.
35
00:02:01,040 --> 00:02:01,820
What school is that.
36
00:02:01,820 --> 00:02:04,160
We actually have Olias coming in here.
37
00:02:04,280 --> 00:02:07,780
We have all kinds of layers tool is here Max pooling.
38
00:02:07,840 --> 00:02:13,050
It shows input outputs what dropout does doesn't change a thing just drops out some layers.
39
00:02:13,250 --> 00:02:19,330
While training Flaten which we know what it does now are dense connections here or drop out again and
40
00:02:19,330 --> 00:02:21,130
are fully connectedly the end here.
41
00:02:21,200 --> 00:02:23,390
Which is also for the connectedly here.
42
00:02:23,930 --> 00:02:25,260
And basically this is it.
43
00:02:25,340 --> 00:02:30,190
So if you were to go to this directory here let's go to the planning directory.
44
00:02:30,290 --> 00:02:32,110
Let's go to train models.
45
00:02:32,330 --> 00:02:41,420
We can see here it is and it's clearer and sharper than I in the book being a DNG file and that's it.
46
00:02:41,420 --> 00:02:43,130
So we've just successfully saved.
|