1 00:00:01,440 --> 00:00:09,180 Here we can create another neural nets for train, the same neural net for the second target that we 2 00:00:09,180 --> 00:00:09,580 have. 3 00:00:10,290 --> 00:00:18,750 But if the targets are having somehow a relation and we can train the same neural network for boat, 4 00:00:19,110 --> 00:00:22,650 then why creating another one here? 5 00:00:22,650 --> 00:00:28,590 I'm going to import all the targets I want to show them separate. 6 00:00:28,590 --> 00:00:35,420 In a separate figures, I use the same neural network to train for Boult outputs. 7 00:00:35,790 --> 00:00:44,120 Again, I can't do that because in this particular examples, the old foods are somehow having a relation. 8 00:00:44,430 --> 00:00:49,550 That's why I can use the same neural network for training both targets. 9 00:00:49,680 --> 00:00:54,870 But if your targets are totally different, then you need to train a separate neural network. 10 00:00:55,470 --> 00:01:01,310 Now, the endpoint point at all the targets, all the two targets in the same neural network. 11 00:01:01,620 --> 00:01:08,940 Let's show the results for the separate outputs to do that copy's part. 12 00:01:08,940 --> 00:01:10,530 Paschi one more time. 13 00:01:10,530 --> 00:01:18,120 The same this one articulates for first targets. 14 00:01:18,510 --> 00:01:25,110 And this one is all the plot for second target. 15 00:01:28,030 --> 00:01:36,370 To specify this target, I need to open and close apprentice's and choose the first target, do the 16 00:01:36,370 --> 00:01:48,390 same for this one and choose the second one, you need to copy this hard and pasted for all the network 17 00:01:48,400 --> 00:01:54,880 units out food for that train target, train out foods. 18 00:01:55,420 --> 00:01:58,630 Do the same for the second target here. 19 00:01:58,630 --> 00:02:07,300 I'm going to copy and paste it for all of them to see the result for the second output. 20 00:02:08,140 --> 00:02:11,140 Of course, change the name to specify them. 21 00:02:11,140 --> 00:02:15,580 This one is for the first one and this is for the second one. 22 00:02:17,410 --> 00:02:25,090 Now, if I were on the program, I should be able to see eight different figures in eight different 23 00:02:25,090 --> 00:02:31,050 windows, four of them are for the first output and four of them are for the second output. 24 00:02:31,450 --> 00:02:34,150 So let's see the result. 25 00:02:37,430 --> 00:02:41,400 OK, it already heated a number of Pook 500. 26 00:02:41,420 --> 00:02:47,850 Maybe the reason is because we have too many old boots we need to increase, maybe to 1000. 27 00:02:48,230 --> 00:02:49,970 Now, you have eight figures. 28 00:02:50,660 --> 00:02:59,060 This one is to see the second output result, for example, for the test data. 29 00:03:00,470 --> 00:03:03,710 Let's compare to test data with this test data. 30 00:03:03,950 --> 00:03:09,380 This is a test data result for the first one, for the first output. 31 00:03:09,380 --> 00:03:12,580 And this is a test data result for the second output. 32 00:03:13,070 --> 00:03:20,390 As you can see, the 15 networks defeating RAF's are a bit different because we have totally different 33 00:03:20,390 --> 00:03:21,190 outputs. 34 00:03:22,280 --> 00:03:29,060 So if you are writing a paper, you can simply train your network using these techniques and then show 35 00:03:29,060 --> 00:03:34,220 the result with different figures, which you really are looking for. 36 00:03:34,220 --> 00:03:42,860 Training data and for testing data can have a figure for training and another separate figure for testing. 37 00:03:43,280 --> 00:03:50,480 Validation is not very over interest, but if you can show it, that's better and just analyze it, 38 00:03:50,480 --> 00:03:56,750 comparing different out foods and complete the training process. 39 00:03:57,110 --> 00:04:01,790 OK, here is all you need to know for training and neural network. 40 00:04:01,790 --> 00:04:08,350 You can just load different samples and adjust it to have the best results.