1
00:00:00,900 --> 00:00:09,060
Here we have an option for retraining and that for using the same data, but it will train the network

2
00:00:09,060 --> 00:00:09,460
again.

3
00:00:09,870 --> 00:00:12,940
So let's just click on it and see what would happen here.

4
00:00:13,500 --> 00:00:20,780
Again, the validation check is stopped over training, but we have different performance purities.

5
00:00:20,790 --> 00:00:23,870
This is a start of failing and failing.

6
00:00:23,880 --> 00:00:28,210
And after six times, the process was already a setup.

7
00:00:28,980 --> 00:00:30,790
Let's check the training estate.

8
00:00:31,050 --> 00:00:38,030
OK, this one is over training estate and this is where the foundation is start to failing.

9
00:00:39,720 --> 00:00:43,500
Let's check the regression testing.

10
00:00:43,500 --> 00:00:45,030
This time is actually better.

11
00:00:46,680 --> 00:00:47,760
I'm not very good.

12
00:00:47,850 --> 00:00:50,940
Not very, but better than the previous time.

13
00:00:51,330 --> 00:00:54,750
And here, just try it again and we can retrain.

14
00:00:54,750 --> 00:00:57,690
You will see different output.

15
00:00:58,380 --> 00:01:00,210
Let me check the performance.

16
00:01:01,020 --> 00:01:03,590
OK, here is a different performance.

17
00:01:03,600 --> 00:01:06,420
And finally, how can it fit?

18
00:01:06,450 --> 00:01:10,040
OK, here is the fitting sample.

19
00:01:11,220 --> 00:01:15,680
So you should actually expect a very different result.

20
00:01:15,690 --> 00:01:22,440
But however, since it's choosing the data randomly, meaning each time it would be a bit different.

21
00:01:23,670 --> 00:01:32,520
Now the team that I want to do is give it more samples and see how can our network behave if it has

22
00:01:32,640 --> 00:01:33,840
more samples.