WEBVTT 00:00.080 --> 00:01.040 Welcome back. 00:01.080 --> 00:08.000 Let's use this validation errors composable in our car features form. 00:08.000 --> 00:13.720 And here you see that we didn't use this validation errors parameter. 00:13.920 --> 00:19.920 Scroll down to under the origin card origin chip. 00:20.080 --> 00:26.680 And under the scope of this row let me use the validation errors. 00:26.680 --> 00:35.760 If validation errors is not empty, go and display the validation errors composable with the validation 00:35.760 --> 00:37.000 errors inside it. 00:37.200 --> 00:47.000 And also you can make a spacer between the validation errors and the button or the row that here of 00:47.280 --> 00:48.280 origin chip. 00:48.320 --> 00:50.760 Okay so let me run. 00:50.760 --> 01:00.360 So if the user missed anything here, there is an error displaying that you should specify the correct 01:00.360 --> 01:04.000 input or you should specify the missing inputs. 01:04.040 --> 01:06.800 Now I want from you to focus with me. 01:06.840 --> 01:08.000 We finished the UI. 01:08.360 --> 01:13.400 Now let's move to the functional coding here. 01:13.560 --> 01:18.320 Create a new package named as predictions. 01:18.320 --> 01:22.600 And inside this package, create a new Kotlin class. 01:22.800 --> 01:25.040 Name it as car. 01:25.160 --> 01:27.320 My mileage predictor. 01:27.360 --> 01:36.080 We need to connect our user interface with the TensorFlow Lite model that we created. 01:36.240 --> 01:38.680 We need to get those data. 01:38.680 --> 01:45.800 And when the user clicks on the submit button, we need to get the data into TensorFlow. 01:46.040 --> 01:53.160 And the TensorFlow will predict the mileage and display the result here. 01:53.280 --> 01:56.280 So this is a connection. 01:56.280 --> 02:02.560 This is a bridge between our TensorFlow and the user interface. 02:02.600 --> 02:11.630 So I want from you to understand it like it's a bridge between TensorFlow model And user interface. 02:11.630 --> 02:18.630 This car mileage predictor class takes into the constructor a context object. 02:18.830 --> 02:24.670 And as we did before, let's start with the interpreter. 02:24.830 --> 02:25.910 Interpreter. 02:26.070 --> 02:29.630 It's from TensorFlow Lite. 02:29.950 --> 02:33.790 It's of type TensorFlow Lite interpreter. 02:33.910 --> 02:39.310 This line of code declares a private variable named interpreter. 02:39.670 --> 02:47.310 That is your app engine for running TensorFlow Lite machine learning models. 02:47.350 --> 02:48.270 TF Lite. 02:48.510 --> 02:57.230 So you give it a model and some input data and it gives you back the predictions or results. 02:57.270 --> 02:57.830 Okay. 02:58.030 --> 03:02.110 The second step is to create the input shape and the output shape. 03:02.230 --> 03:06.070 This is very important in machine learning. 03:06.110 --> 03:12.670 We need to get the shape of our input and get the output shape of our model. 03:12.710 --> 03:16.830 Private val input shape. 03:16.870 --> 03:20.310 What is the shape of our model? 03:20.470 --> 03:25.190 It's an int array of one and nine. 03:25.350 --> 03:29.950 We need to pass one sample and nine features. 03:29.990 --> 03:31.630 Did you remember the features? 03:31.630 --> 03:33.110 Let me go back to here. 03:33.150 --> 03:38.910 Car features 123, four, five, six, seven, eight and nine. 03:39.070 --> 03:48.910 So we need to set the input shape to 1999 features and one sample. 03:49.110 --> 03:55.710 So one sample and nine features we need to get the output shape. 03:55.710 --> 04:03.270 So output shape equals to int array of one of one. 04:03.430 --> 04:06.750 So here we have one sample. 04:06.750 --> 04:10.990 And we need to get only the mileage. 04:11.150 --> 04:13.990 So the output would be a one number. 04:14.150 --> 04:16.870 It's the mileage of our car. 04:16.910 --> 04:17.390 Okay. 04:17.590 --> 04:19.940 So the output shape. 04:19.980 --> 04:23.300 Output shape is a number. 04:23.340 --> 04:25.980 This is very very important. 04:26.220 --> 04:32.420 We need to understand the input and the output of our shape and our model. 04:32.460 --> 04:32.940 Okay. 04:33.100 --> 04:35.940 So give me nine parameters. 04:35.980 --> 04:41.340 Nine features I'll give you one output okay. 04:41.700 --> 04:46.380 So this is the meaning of the input and output shape here. 04:46.420 --> 04:47.260 To solve it. 04:47.300 --> 04:48.380 We need to use it. 04:48.420 --> 04:52.060 Private var interpreter add late. 04:52.060 --> 04:52.660 Initiate. 04:52.660 --> 04:54.780 You can late initiate it okay. 04:54.940 --> 04:58.220 Then start with init block. 04:58.260 --> 05:01.180 We need to load that Tflite model. 05:01.180 --> 05:09.620 So here val model equals to load model file and set the context. 05:09.820 --> 05:17.100 The parameter interpreter equals to interpreter and pass the model. 05:17.140 --> 05:20.620 In the next video we're going to create this function.