1
00:00:00,900 --> 00:00:07,740
The last thing that we have to do is drain our model training model is usually done using the fixed

2
00:00:07,740 --> 00:00:15,330
function, but this time we are going to use deep fixed generator function in this big generator function.

3
00:00:15,630 --> 00:00:19,260
We can give Grangemouth that as one parameter.

4
00:00:21,240 --> 00:00:29,340
So this big generator function will get input of 20 images at a time from our range Endacott function.

5
00:00:31,110 --> 00:00:38,280
Does it function that we created a loop to destroy Ingenico, which of which was giving out 20 images

6
00:00:38,370 --> 00:00:40,560
of 150 where one of PDM engines.

7
00:00:43,670 --> 00:00:46,850
That is a pattern we got put beginning to function.

8
00:00:49,630 --> 00:00:53,590
Now we have to specify this steps, but Epoch Barometer.

9
00:00:53,960 --> 00:00:54,950
This is mandatory.

10
00:00:56,450 --> 00:01:02,990
It is very important because as I told you earlier, that train, that is an infinite loop.

11
00:01:03,410 --> 00:01:06,920
It will keep on giving you and the images continuously.

12
00:01:07,940 --> 00:01:09,680
You have to tell when to stop.

13
00:01:10,670 --> 00:01:12,590
So steps but epoch.

14
00:01:13,160 --> 00:01:17,630
The eldest ranger needs to stop after giving the images.

15
00:01:17,830 --> 00:01:18,680
Hunger pains.

16
00:01:19,820 --> 00:01:20,630
So steps.

17
00:01:20,660 --> 00:01:24,410
But epoch basically means that within one epoch.

18
00:01:25,790 --> 00:01:28,280
How many times are we learning this.

19
00:01:28,280 --> 00:01:28,980
Strange that.

20
00:01:31,170 --> 00:01:41,010
Since we were training on a data set of 2000 images, we wanted to input 2000 images into what pretended

21
00:01:41,040 --> 00:01:44,600
to danger Nader was giving us going images.

22
00:01:44,730 --> 00:01:50,570
But that so to train on 2000 images.

23
00:01:51,240 --> 00:01:58,140
We done that Prange Anuta a hundred times so that we get 2000 training images in different generative

24
00:01:58,140 --> 00:01:58,530
function.

25
00:02:01,170 --> 00:02:02,660
Next parameters, epochs.

26
00:02:02,900 --> 00:02:08,040
This is the number of planes, these two images will be fed into our model.

27
00:02:08,780 --> 00:02:12,420
We have specified it to grindy validation.

28
00:02:12,420 --> 00:02:19,590
Data is to be taken form from validation generator and validation steps are set to complete.

29
00:02:20,220 --> 00:02:23,040
Because this is also giving grindy images.

30
00:02:24,460 --> 00:02:26,350
So 15 to 20 is 1000.

31
00:02:26,880 --> 00:02:30,090
And we have a data of 1000 validation images.

32
00:02:31,350 --> 00:02:36,150
So that is why steps but epoch in training is hundred and invalidation steps.

33
00:02:36,260 --> 00:02:36,850
It is pithy.

34
00:02:39,530 --> 00:02:40,480
Bananas on this.

35
00:02:41,240 --> 00:02:47,630
This is going to take a lot of pain because we have a Bigfoot model, this thing to be prepared.

36
00:02:47,820 --> 00:02:49,310
It may take up to half an hour.
