1 00:00:00,490 --> 00:00:02,860 Now after we. 2 00:00:03,880 --> 00:00:05,200 It will define the network. 3 00:00:05,200 --> 00:00:07,300 We define the initial initial condition. 4 00:00:07,300 --> 00:00:14,680 It will be random, the boundary condition and of course, why it's initial condition we didn't care 5 00:00:14,680 --> 00:00:17,070 about because it's a steady state. 6 00:00:17,080 --> 00:00:19,660 If we want to define the initial condition, this is okay. 7 00:00:20,020 --> 00:00:25,240 But because it's a steady state problem, we don't really care if it's just defined as a random. 8 00:00:25,690 --> 00:00:33,670 So in order to train the network, what we need to do is we start by model defining the model. 9 00:00:36,780 --> 00:00:39,660 We want to d e. 10 00:00:40,920 --> 00:00:42,030 Dot model. 11 00:00:42,120 --> 00:00:45,780 It will have the data and will have the network. 12 00:00:45,780 --> 00:00:46,140 This one. 13 00:00:46,170 --> 00:00:49,100 Of course, the data is this one. 14 00:00:49,110 --> 00:00:53,400 What we defined here and the network is this one. 15 00:00:55,050 --> 00:00:55,950 And. 16 00:00:56,880 --> 00:00:57,630 Model. 17 00:00:58,560 --> 00:01:00,750 Dot com compile. 18 00:01:02,750 --> 00:01:06,800 Pyle and the first Adam. 19 00:01:06,800 --> 00:01:09,590 We will use the same methods. 20 00:01:09,590 --> 00:01:10,250 Both. 21 00:01:11,410 --> 00:01:13,360 Knee to the polls, minus three. 22 00:01:14,840 --> 00:01:17,960 Shift into a mode. 23 00:01:23,970 --> 00:01:26,220 Uh, this is a capital letter. 24 00:01:27,890 --> 00:01:28,040 More. 25 00:01:28,510 --> 00:01:29,030 Yeah. 26 00:01:29,030 --> 00:01:33,650 Now is okay so this one now. 27 00:01:36,210 --> 00:01:40,470 We will have lost history. 28 00:01:41,970 --> 00:01:42,860 History. 29 00:01:43,500 --> 00:01:45,300 Train State. 30 00:01:46,200 --> 00:01:53,040 And model dot train with epochs of. 31 00:01:53,870 --> 00:01:54,800 10,000. 32 00:01:54,830 --> 00:01:58,100 It will take this time a little bit longer than before. 33 00:01:58,430 --> 00:02:01,310 And we need a. 34 00:02:02,960 --> 00:02:03,920 This is the same thing. 35 00:02:03,920 --> 00:02:09,200 Like, we will start with Adam and after Adam we need to do the other method. 36 00:02:09,200 --> 00:02:10,070 So. 37 00:02:11,310 --> 00:02:12,120 I'll just. 38 00:02:12,120 --> 00:02:13,410 I will not start this one. 39 00:02:13,440 --> 00:02:19,020 It will take like the e dot of to. 40 00:02:23,920 --> 00:02:25,720 And this is compiled already. 41 00:02:25,720 --> 00:02:30,430 Start up my Zeus dot. 42 00:02:32,720 --> 00:02:34,400 Config. 43 00:02:38,530 --> 00:02:40,480 Dot set. 44 00:02:41,570 --> 00:02:42,140 L. 45 00:02:42,140 --> 00:02:42,700 P. 46 00:02:42,710 --> 00:02:43,400 F. 47 00:02:43,400 --> 00:02:43,860 G. 48 00:02:44,330 --> 00:02:44,520 L. 49 00:02:44,570 --> 00:02:44,900 P. 50 00:02:46,370 --> 00:02:50,570 If these options. 51 00:02:51,600 --> 00:02:54,570 This is just max iteration. 52 00:02:55,460 --> 00:02:57,500 To set basically the max iteration. 53 00:02:57,500 --> 00:02:58,640 3000. 54 00:02:59,600 --> 00:03:01,310 Max iteration. 55 00:03:02,180 --> 00:03:06,200 And model dot compile. 56 00:03:08,510 --> 00:03:19,310 And slash bfgs the same method before and we need to do the same here. 57 00:03:20,430 --> 00:03:25,320 But we don't have to consider this one because we already define it here. 58 00:03:26,550 --> 00:03:27,210 Epok. 59 00:03:27,210 --> 00:03:27,930 No need. 60 00:03:28,380 --> 00:03:29,400 And. 61 00:03:31,590 --> 00:03:32,850 Yeah, we can save. 62 00:03:32,850 --> 00:03:34,380 Like the values. 63 00:03:34,410 --> 00:03:36,690 Dot save. 64 00:03:37,880 --> 00:03:38,570 Plot. 65 00:03:41,110 --> 00:03:43,840 Uh, just lost history and strained. 66 00:03:47,700 --> 00:03:51,630 Nisei save equals false. 67 00:03:54,160 --> 00:03:54,820 Pulse. 68 00:03:56,010 --> 00:03:57,930 Is plot. 69 00:03:59,560 --> 00:04:02,500 Because I don't really want to plot them. 70 00:04:04,100 --> 00:04:04,550 Yeah. 71 00:04:05,150 --> 00:04:05,600 Okay. 72 00:04:05,600 --> 00:04:11,660 So this is basically we defined the model, we compile it using Adam. 73 00:04:11,660 --> 00:04:14,840 We will train it once we push this one shift enter. 74 00:04:14,840 --> 00:04:25,940 It's now training and after that training it will go to optimizers and will um, well basically use 75 00:04:25,940 --> 00:04:34,420 the limited memory this, this one LPF optimizer and it will have a. 76 00:04:35,600 --> 00:04:39,380 Maximum iteration of 3000 and that's it. 77 00:04:39,380 --> 00:04:41,690 And shift enter after this. 78 00:04:41,720 --> 00:04:44,000 It will train with that. 79 00:04:44,000 --> 00:04:48,560 We already finished training our module. 80 00:04:50,000 --> 00:04:51,590 Well, that's it. 81 00:04:52,010 --> 00:04:53,270 Let's wait. 82 00:04:53,270 --> 00:05:01,160 And once it's finished, we will go to the post-processing and we will see our results that hopefully 83 00:05:01,160 --> 00:05:02,480 it will converge.