WEBVTT 00:00.080 --> 00:00.920 Welcome back. 00:00.960 --> 00:07.960 Ultralytic supports a wide range of YOLO models, from early versions like YOLO version three to the 00:07.960 --> 00:10.880 latest YOLO version 11 and 12. 00:10.920 --> 00:21.200 Okay, so there are a lot of ways you can use YOLO for detecting segmentation, classification, pose 00:21.200 --> 00:25.960 detection, object detection, and tracking. 00:26.200 --> 00:34.640 Okay, there are a lot of types and uh, there are a lot of models featuring 80 object classes. 00:34.640 --> 00:37.400 So this is that pre-trained model. 00:37.600 --> 00:43.440 The more these models are trained on the Coco data set, you remember that Coco data set that we talked 00:43.440 --> 00:46.280 about in the previous videos. 00:46.440 --> 00:56.480 And this Coco dataset contains 80 objects from person from uh, a car, truck and others. 00:56.520 --> 00:57.000 Okay. 00:57.240 --> 00:59.720 But we are interested in those models. 00:59.720 --> 01:04.920 We have YOLO, YOLO 11 and N, S, M, L and x. 01:05.160 --> 01:06.440 This is nano. 01:06.600 --> 01:10.920 This is small size and can detect up to um. 01:11.280 --> 01:20.960 If the speed is milliseconds and the parameters is point to point 6 million, the YOLO 11 s, which 01:20.960 --> 01:25.880 is small, the size in pixels 640 by 640. 01:25.920 --> 01:27.080 This is the image size. 01:27.080 --> 01:35.160 And with the height, the width and height and medium is uh, bigger, large is is larger. 01:35.160 --> 01:41.160 And Xlarge is the largest model you can use for Android. 01:41.480 --> 01:46.240 It's recommended to use one of those N or S. 01:46.280 --> 01:46.720 Okay. 01:46.760 --> 01:55.200 So those are the models that we're going to use in our object detection with androids with Android applications. 01:55.240 --> 01:55.760 Okay. 01:55.800 --> 02:01.280 Now let's go back to our YOLO custom object detection file. 02:01.280 --> 02:04.800 And here from Roboflow import or overflow. 02:05.040 --> 02:14.400 This is the roboflow works or um repository can name it as repository and passing the API key and creating 02:14.440 --> 02:17.800 an object from this roboflow. 02:17.800 --> 02:23.960 Passing the API key the project get from Roboflow the workspace object detection. 02:23.960 --> 02:27.160 By the way, don't share this API key with anyone. 02:27.160 --> 02:28.480 Keep it secret with your. 02:28.680 --> 02:31.920 But this is for used for demonstration, so don't worry. 02:31.920 --> 02:34.680 You can, um, you can uh. 02:35.000 --> 02:38.720 This is why I'm sharing with you the object detection. 02:38.720 --> 02:41.480 And by the way, this is the API is different from yours. 02:41.520 --> 02:42.040 Okay? 02:42.120 --> 02:44.200 So don't use this API key. 02:44.240 --> 02:45.880 Use your own API key. 02:45.920 --> 02:53.920 As I showed you in the previous videos, the flow dot workspace object detection and the project is 02:53.960 --> 02:55.240 this version. 02:55.240 --> 03:00.760 Get from the project version number one and data set equals to version dot download. 03:00.760 --> 03:06.190 You see guys I'm using YOLO version eight okay now click. 03:06.190 --> 03:06.870 Run. 03:06.910 --> 03:08.350 Open the folder. 03:08.510 --> 03:10.790 You see we have only sample data. 03:10.990 --> 03:20.630 Wait for collecting and downloading all the data set from here and you will see the newly created folder. 03:20.630 --> 03:21.470 And here we go. 03:21.510 --> 03:25.510 This is the newly created folder called Object Detection Android one. 03:25.510 --> 03:35.070 If we go to our roboflow you see guys that um, that our object object detection, object detection 03:35.070 --> 03:37.070 Android, this is our project. 03:37.070 --> 03:39.670 It's named as object detection okay. 03:39.710 --> 03:43.630 And the version if we go to versions, you see version one. 03:43.630 --> 03:50.910 For this we get the, uh, this, um, this naming object detection Android one. 03:51.070 --> 03:57.550 If we open it, we have many files data, dot, YAML, ui, ml. 03:57.590 --> 04:01.070 This is the most important file that we're going to use. 04:01.270 --> 04:11.590 And here we have three folders test train and valid Inside every folder there are two folders images 04:11.590 --> 04:12.470 and labels. 04:12.470 --> 04:20.590 If we go to the images, we see images, and if we go to the labels, we see the labeled ones and the 04:20.590 --> 04:21.190 labels. 04:21.230 --> 04:23.750 Okay dot, txt open the txt. 04:23.790 --> 04:27.390 You see those are the bounding boxes of the detected object. 04:27.390 --> 04:33.230 And if we go to the photos you see we have this photo for example. 04:33.510 --> 04:35.270 This is another photo. 04:35.390 --> 04:41.670 This is the uh the the images that we uploaded to our Roboflow images. 04:41.670 --> 04:43.310 Those are the test. 04:43.390 --> 04:47.710 You can see the train and labels also you see valid. 04:47.710 --> 04:50.710 We have images and labels okay. 04:50.950 --> 04:52.750 Those are very important. 04:52.750 --> 05:00.670 We're going to use those images to train our model using Python and Google Colab with the help of YOLO 05:00.950 --> 05:02.510 object detections okay.