WEBVTT 00:00.080 --> 00:01.040 Welcome back. 00:01.080 --> 00:03.520 We finished the object detector class. 00:03.520 --> 00:06.680 Now let's create the view model. 00:06.680 --> 00:14.400 Inside this package, create a new package called View Model and inside it create a new Kotlin class. 00:14.400 --> 00:18.200 Name it as Detection View model. 00:18.240 --> 00:25.880 If you remember, we added Hilt Dependency Injection library to our Gradle file. 00:26.120 --> 00:30.440 So we're going to use here to do the dependency injection. 00:30.440 --> 00:37.080 And as we learned before the view model will be marked as hilt view model. 00:37.120 --> 00:44.000 Please go back to the dependency injection section in order to learn more about the hilt view model. 00:44.040 --> 00:54.720 Also we need to inject the constructor with inject annotation and here we need the private val context 00:54.720 --> 01:03.560 context and this dependency can be provided by hilt by using another annotation called Application Context, 01:03.560 --> 01:10.280 and to tell Android Studio that this class will act as a ViewModel, we need to extend from ViewModel. 01:10.320 --> 01:19.040 Now let's create a companion object, which is which is containing the constants that we need to use 01:19.040 --> 01:20.600 companion object. 01:20.600 --> 01:27.240 And here let me start with the constant val tag equals to tag. 01:27.440 --> 01:28.480 This is for luck. 01:28.520 --> 01:33.520 Also we need to add the model file name and label file name. 01:33.560 --> 01:44.200 Let's start with constant val model file name and the constant val label file name. 01:44.200 --> 01:52.400 Let's add the model that we, uh, we downloaded from SSD MobileNet and from Kaggle website. 01:52.440 --> 01:54.080 To add it, it's very simple. 01:54.080 --> 02:00.280 Inside this app, create a new other TensorFlow Lite model. 02:00.400 --> 02:03.670 Click on it, Locate your model. 02:03.990 --> 02:06.590 Select the flight automatically. 02:06.630 --> 02:09.590 Add build feature and dependencies to build your Gradle. 02:09.630 --> 02:12.430 Keep it like this and click finish. 02:12.470 --> 02:17.430 Wait for Android Studio to add the model to the ML file and folder. 02:17.430 --> 02:18.830 And this is our model. 02:19.030 --> 02:28.630 If you notice that we have an int array of one 300 303, data type is unsigned integer eight. 02:28.870 --> 02:29.430 Okay. 02:29.470 --> 02:32.590 We talked about the shape of our model. 02:32.790 --> 02:36.630 And this is just a confirmation of the inputs shapes. 02:36.670 --> 02:37.070 Okay. 02:37.110 --> 02:38.390 So this is our model. 02:38.430 --> 02:43.750 Add it inside the ML package that is created by Android Studio. 02:43.830 --> 02:46.870 The second step is the label file. 02:47.070 --> 02:52.830 Go to Google search for download Coco dataset labels version one dot txt. 02:52.910 --> 02:55.150 Here we have the annotations. 02:55.150 --> 02:57.830 The train 2017. 02:57.870 --> 02:58.710 It's similar. 02:59.070 --> 03:07.550 You can go to Coco dataset to download the whole data sets, but there are a lot of libraries that provide 03:07.550 --> 03:08.750 you with the labels. 03:08.750 --> 03:10.150 You can go to GitHub. 03:10.190 --> 03:13.430 Me leave labels. 03:13.470 --> 03:22.590 I clicked on this repository and you can download the Coco labels 2014 2017. 03:22.630 --> 03:24.150 Those are the labels. 03:24.270 --> 03:30.310 Copy them or you can download them as a row like this copy. 03:30.550 --> 03:33.910 And here we need to create the text file. 03:33.910 --> 03:36.390 Create a new text document. 03:36.430 --> 03:44.230 Name it as Coco data, set labels, open it and paste those labels. 03:44.230 --> 03:45.510 Click save. 03:45.630 --> 03:46.790 And here we go. 03:46.830 --> 03:49.030 This is the Coco data set labels. 03:49.030 --> 03:54.510 Let's go to Android Studio and copy this file into the folder. 03:54.590 --> 03:56.430 Paste and click okay. 03:56.550 --> 04:02.740 Those are the labels starting from person and ending with tooth brush. 04:02.780 --> 04:06.540 Okay, now let's add the model file name. 04:06.540 --> 04:15.140 Here we have, uh, the Coco data set labels and mobile as the SSD MobileNet version one Tflite. 04:15.180 --> 04:19.620 Please pay attention for the name so it's case sensitive. 04:19.860 --> 04:25.780 It's called SSD underscore MobileNet underscore version one TF Lite. 04:25.820 --> 04:29.380 Make sure that it is um highlighted by green. 04:29.380 --> 04:31.860 So you hit the correct file. 04:31.860 --> 04:35.060 Then use the label file name. 04:35.220 --> 04:37.700 The label file name is Coco. 04:37.740 --> 04:40.380 This is the name Coco. 04:40.420 --> 04:46.180 Underscore data set underscore labels dot txt okay. 04:46.380 --> 04:52.420 It's highlighted by green and it's hitting the correct file okay. 04:52.460 --> 05:00.060 In this way we added the SSD mobile version one tflite file and the Coco dataset labels.