1
00:00:02,810 --> 00:00:09,650
Yolo v nine is the new computer vision object detection model released by Jin Yong Wang and his team

2
00:00:09,650 --> 00:00:12,170
on 21st February 2024.

3
00:00:12,200 --> 00:00:20,210
In this tutorial, we will review how we can do object detection on images and on videos using YOLO

4
00:00:20,210 --> 00:00:22,190
v9 in Google Colab.

5
00:00:22,760 --> 00:00:28,670
Before you run this script, please make sure that you have selected the runtime as T4 GPU.

6
00:00:28,700 --> 00:00:32,060
Google Colab offers free GPU so you can use this.

7
00:00:32,420 --> 00:00:37,340
So in the first step, we will clone this yolo v9 GitHub repo.

8
00:00:37,340 --> 00:00:41,780
So you can simply go here and you can just click copy this URL from here.

9
00:00:41,930 --> 00:00:45,770
And you can just come back to the Google Colab notebook.

10
00:00:45,770 --> 00:00:48,050
And over here you will write git clone.

11
00:00:48,050 --> 00:00:51,380
And here you will pass the URL that you have copied.

12
00:00:51,380 --> 00:00:54,590
So here you will just add the URL that you have copied.

13
00:00:54,890 --> 00:01:01,010
So now we will just run this cell and we will be able to clone the GitHub repo into our Google Colab

14
00:01:01,010 --> 00:01:01,430
notebook.

15
00:01:01,430 --> 00:01:05,060
And here you can see the complete repo which we have cloned.

16
00:01:06,560 --> 00:01:11,300
So Golovnin introduced two new architectures, Golovnin and Glenn.

17
00:01:11,300 --> 00:01:15,200
So in the Yolo v nine paper, two new architectures have been introduced.

18
00:01:15,200 --> 00:01:18,080
One is the Yolo v nine and other is the Glenn.

19
00:01:18,560 --> 00:01:23,690
Both model weights are available in the Yolo v nine repository released with the paper.

20
00:01:23,690 --> 00:01:25,190
So if you just go over here.

21
00:01:25,190 --> 00:01:32,420
So and if you just click on this version 0.1 and here you will find the Glenn model weights as well

22
00:01:32,990 --> 00:01:38,630
okay Glenn C I Glenn E and yolo V9C Yolo V9E.

23
00:01:38,630 --> 00:01:43,070
So you will find the Glenn model weights as well as the Yolo v nine model weights.

24
00:01:43,070 --> 00:01:50,600
So YOLO v nine and Glenn come with four different models 1V9 small, YOLO nine compact, YOLO v nine

25
00:01:50,600 --> 00:01:52,730
medium and Yolo v nine extended.

26
00:01:54,460 --> 00:02:00,550
So if you just go over here, you will find the list of models that YOLO v9 come with YOLO v9, small,

27
00:02:00,550 --> 00:02:04,090
YOLO v9 medium, YOLO v9 compact, and YOLO v9 extended.

28
00:02:04,090 --> 00:02:08,530
Similarly, you can also come with all these models as well.

29
00:02:10,670 --> 00:02:17,090
So in this tutorial we will see how we can do object detection with YOLO v9 as well as with Glen model

30
00:02:17,090 --> 00:02:17,750
as well.

31
00:02:21,350 --> 00:02:25,490
So in the next step we will download the pre-trained model weights from the GitHub repo.

32
00:02:25,520 --> 00:02:31,850
The weights for the Yolov5 small model and Golovnin medium model are not available at the time of writing

33
00:02:31,850 --> 00:02:32,420
this notebook.

34
00:02:32,420 --> 00:02:38,660
So if you just see over here, uh, yolo v9 comes with four different models YOLO v9, small, medium,

35
00:02:38,660 --> 00:02:39,920
compact and extended.

36
00:02:39,920 --> 00:02:45,110
But over here you will see that we have only weights available for the compact model as well as for

37
00:02:45,110 --> 00:02:46,280
the extended model.

38
00:02:46,280 --> 00:02:51,740
And similarly in the case of Glenn, we have weights available for the compact model as well as for

39
00:02:51,740 --> 00:02:52,760
the extended model.

40
00:02:52,760 --> 00:02:57,620
We have don't have the weights available for the small model or for the medium model.

41
00:02:57,620 --> 00:03:01,760
So we have two weight model, uh, weights available, compact and extended.

42
00:03:01,760 --> 00:03:06,110
While we don't have the weights available for the small or the medium model.

43
00:03:07,150 --> 00:03:13,030
So to download this weights, you can simply, uh, click over here, copy the link address and you

44
00:03:13,030 --> 00:03:14,890
can simply write w get.

45
00:03:14,890 --> 00:03:17,320
And you can add the link address over here.

46
00:03:17,320 --> 00:03:22,570
And you will be able to download all these weights into your Google Colab notebook.

47
00:03:22,570 --> 00:03:25,600
So I will just download all the weights.

48
00:03:25,600 --> 00:03:32,290
So I'm just downloading the compact and extended weights for the Glen and YOLO v9 model.

49
00:03:32,890 --> 00:03:38,200
So here you can see that I have downloaded Golovnin compact model weights, YOLO v9 extended model weights.

50
00:03:38,200 --> 00:03:43,750
And similarly here I've downloaded Glenn compact model weights and Glenn Extended model weights.

51
00:03:46,810 --> 00:03:53,170
So now in the next step, we will set our Golovnin folder, which we have cloned over here, our Yolov5

52
00:03:53,170 --> 00:03:57,580
repository, which we have cloned into our Google Colab notebook.

53
00:03:57,580 --> 00:04:04,480
So I will just set this, uh, directory or uh, I will just set, uh, change my directory to yolo

54
00:04:04,480 --> 00:04:05,080
v nine.

55
00:04:06,270 --> 00:04:09,180
So I've just changed the directory to YOLO v nine folder.

56
00:04:09,180 --> 00:04:12,030
So now we will install all the dependencies.

57
00:04:12,120 --> 00:04:13,800
If you just see the requirements dot.

58
00:04:13,800 --> 00:04:16,260
TXT file in the requirements dot txt file.

59
00:04:16,260 --> 00:04:22,800
We have listed all the requirements or all the libraries that are required to run the uh to do detection

60
00:04:22,800 --> 00:04:24,270
on images and videos.

61
00:04:24,270 --> 00:04:30,150
So if you want to install all these packages once, you can simply write pip install minus r requirements

62
00:04:30,150 --> 00:04:30,900
dot txt.

63
00:04:30,900 --> 00:04:36,840
So it will install all the libraries dependencies that are mentioned in the dot txt file.

64
00:04:38,500 --> 00:04:43,660
So I will just run this cell and it will install all the dependencies that are mentioned in the dot.

65
00:04:43,660 --> 00:04:44,650
TXT file.

66
00:04:46,390 --> 00:04:49,960
So this might take a few seconds before it gets completed.

67
00:04:53,810 --> 00:04:54,080
Nope.

68
00:04:54,080 --> 00:04:58,760
We have installed all the packages that are listed in the requirements.txt file.

69
00:04:59,120 --> 00:05:04,610
Now in the step number five we will see how we can do inference on images and videos.

70
00:05:04,670 --> 00:05:08,480
First we will do inference on image using Galen model.

71
00:05:08,480 --> 00:05:12,650
So this is the best model uh in the Galen series.

72
00:05:14,290 --> 00:05:21,070
So first of all, I will just download a random image from my Google Drive into this Google Colab notebook

73
00:05:21,070 --> 00:05:21,640
over here.

74
00:05:21,640 --> 00:05:27,880
So I will just run this cell and it will download that image from my Google Drive into this Google Colab

75
00:05:27,880 --> 00:05:28,540
notebook.

76
00:05:28,540 --> 00:05:34,600
And over here, if I just go to the YOLO v nine folder and over here you will find the image one dot

77
00:05:34,600 --> 00:05:36,610
jpg which I have downloaded.

78
00:05:37,120 --> 00:05:40,990
Okay, so now I will just write over here python detect.py.

79
00:05:40,990 --> 00:05:45,520
And here I will just pass the weights and which I have downloaded already above.

80
00:05:45,520 --> 00:05:51,880
And here I will just pass the image name in the source and device zero because I have selected the runtime

81
00:05:51,880 --> 00:05:52,900
as GPU.

82
00:05:52,900 --> 00:05:56,890
So if I just like to CPU then I will just write device CPU.

83
00:05:56,890 --> 00:05:57,940
But I am using.

84
00:05:57,940 --> 00:06:03,250
I have selected the runtime as GPU, so I am writing device zero and now I will just run this cell.

85
00:06:03,250 --> 00:06:08,320
So if you want to get familiar with more about this command, you can just go to the GitHub repository.

86
00:06:08,320 --> 00:06:13,930
And over here you will see uh, all the steps are written how to uh, if you want to do inference with

87
00:06:13,930 --> 00:06:15,040
the YOLO v nine model.

88
00:06:15,040 --> 00:06:18,070
So you can write Python detect dash dual dot pi.

89
00:06:18,070 --> 00:06:24,460
And if you want to do inference with the Galen models, you can simply write Python detect dot pi.

90
00:06:25,420 --> 00:06:28,660
So I am doing inference on image using an E model.

91
00:06:28,660 --> 00:06:31,330
So I will just write Python detect.py.

92
00:06:31,330 --> 00:06:35,440
And here I will just pass the weights file name and the source and device.

93
00:06:35,440 --> 00:06:36,820
So this is all you require.

94
00:06:36,820 --> 00:06:39,040
So now I will just run this cell over here.

95
00:06:42,200 --> 00:06:44,600
So the inference take a few seconds.

96
00:06:44,600 --> 00:06:47,960
So it is about to be done.

97
00:06:47,960 --> 00:06:54,590
So now you can see over here we are using Tesla T4 GPU and our results are saved in Runs Detect experiment

98
00:06:54,590 --> 00:06:54,980
folder.

99
00:06:54,980 --> 00:07:00,110
So if you just go to the Yolov5 folder and here we have the runs detect and experiment.

100
00:07:00,110 --> 00:07:02,270
And here is my output image saved.

101
00:07:02,270 --> 00:07:09,020
So I will just first from IPython I will just import the image library and let me display my output

102
00:07:09,020 --> 00:07:09,710
image over here.

103
00:07:09,710 --> 00:07:11,990
So you can simply copy path from here.

104
00:07:13,660 --> 00:07:18,100
You can remove this because you'll be nine is already set as our current directory.

105
00:07:18,100 --> 00:07:20,800
And if you just run this cell over here.

106
00:07:20,800 --> 00:07:23,170
So this will display you the output image.

107
00:07:23,170 --> 00:07:28,780
So now you can see over here uh using Galen Extended Model we are able to detect the first person's

108
00:07:28,780 --> 00:07:30,340
handbag backpack.

109
00:07:30,700 --> 00:07:31,120
Uh.

110
00:07:32,450 --> 00:07:34,880
But we have detected traffic lights as well.

111
00:07:34,880 --> 00:07:38,960
And this is not the surfboard, so this is the wrong direction.

112
00:07:39,170 --> 00:07:44,030
Uh, one there is one wrong direction, else all other directions look pretty fine.

113
00:07:44,030 --> 00:07:48,860
It has, uh, detected persons which are far away or blur as well.

114
00:07:48,860 --> 00:07:53,990
So now we will do inference on the same image using YOLO v9 extended model.

115
00:07:53,990 --> 00:07:54,800
So.

116
00:07:56,120 --> 00:07:57,380
If you just go over here.

117
00:07:57,380 --> 00:07:59,180
So these are the best performing models.

118
00:07:59,180 --> 00:08:00,350
YOLO v9 extended.

119
00:08:00,350 --> 00:08:02,840
They have more number of parameters as well.

120
00:08:02,840 --> 00:08:09,050
And they uh, have good mean average precision as compared to the other YOLO v9 models.

121
00:08:09,050 --> 00:08:15,140
So we are using the best, uh, YOLO v9 model as well as the best Galen model, which is the extended

122
00:08:15,140 --> 00:08:15,830
model.

123
00:08:16,130 --> 00:08:24,560
So now as I told you to do inference on image video using yolo v9, we need to try it over here.

124
00:08:24,560 --> 00:08:28,100
If I just show you we need to write Python detect-dual.py.

125
00:08:28,100 --> 00:08:31,670
So we will write over here python detect-dual.py.

126
00:08:31,670 --> 00:08:34,340
And here I will just pass the weights name.

127
00:08:34,340 --> 00:08:37,100
And here I've selected a source as image one.

128
00:08:37,100 --> 00:08:39,650
I'm using the same image and device as GPU.

129
00:08:39,650 --> 00:08:41,720
For GPU I will write zero.

130
00:08:42,080 --> 00:08:44,450
So now I will just run the cell over here.

131
00:08:48,250 --> 00:08:49,840
So in the inference takes some time.

132
00:08:49,840 --> 00:08:51,220
So now it's using layer.

133
00:08:51,220 --> 00:08:54,970
And you can see that we have uh folder exponential two is being created.

134
00:08:54,970 --> 00:08:57,160
And inside this we have the output image.

135
00:08:57,160 --> 00:09:01,270
And let us uh display this image output image over here.

136
00:09:01,270 --> 00:09:04,240
So you just need to find pass the file path over here.

137
00:09:04,240 --> 00:09:10,060
And you can see that, uh, in the previous, uh, Glenn extended model, we have one wrong detection,

138
00:09:10,060 --> 00:09:15,070
which is sort of word, but, uh, the this is not the case with the yolo v nine over here.

139
00:09:15,070 --> 00:09:19,450
So exactly as yolo v nine outperforms Glenn E model as well.

140
00:09:19,450 --> 00:09:24,100
So now you can see that, uh, we don't have any, uh, false positive over here.

141
00:09:25,350 --> 00:09:31,560
And you can see over here we have the sandbag backpack, a boss person, and it has detected, uh,

142
00:09:31,560 --> 00:09:35,280
persons which are quite far away or blurred, like you can see over here.

143
00:09:35,280 --> 00:09:36,870
So this is good.

144
00:09:38,310 --> 00:09:43,530
Now we will be doing inference on videos using YOLO v9 and Glen model as well.

145
00:09:43,530 --> 00:09:50,700
So let's first download a sample video from the drive into this Google Colab notebook.

146
00:09:50,700 --> 00:09:56,700
I have already placed the sample video onto my drive, and I'm just downloading that video from drive

147
00:09:56,700 --> 00:09:58,590
into this Google Colab notebook.

148
00:09:59,040 --> 00:10:04,110
So now you can see that this demo dot mp4 video is.

149
00:10:06,600 --> 00:10:10,740
Okay, now you can see this demo dot mp4 video over here.

150
00:10:10,740 --> 00:10:11,850
We have downloaded this.

151
00:10:11,850 --> 00:10:15,330
So now I will just try and write Python detect.py.

152
00:10:15,330 --> 00:10:22,290
And here I will just pass the file name and the source as demo dot mp4 and device zero for GPU.

153
00:10:22,290 --> 00:10:23,610
So let's run this cell.

154
00:10:31,560 --> 00:10:37,260
So now you can see that our complete video is being divided into 1314 frames, and the detection on

155
00:10:37,260 --> 00:10:40,710
each of the frame is being done one by one.

156
00:10:41,130 --> 00:10:42,900
So this will take some time.

157
00:10:42,900 --> 00:10:45,090
So let's wait for it to get finished.

158
00:10:45,090 --> 00:10:47,640
And then I will display the output video over here.

159
00:10:49,890 --> 00:10:53,940
So this might take a few more seconds before it gets finished.

160
00:11:14,960 --> 00:11:19,520
So the inference on this input video demo mp4 is being done.

161
00:11:19,520 --> 00:11:22,280
We have used the Glen E model.

162
00:11:22,280 --> 00:11:27,680
So Glen stand for generalized efficient layer aggregation network okay.

163
00:11:27,680 --> 00:11:31,790
So if you I have also displayed the output demo video over here.

164
00:11:31,790 --> 00:11:37,400
So let me just download this from here and let me show you how our output look like.

165
00:11:38,040 --> 00:11:43,770
So here we are using Extended model, the best performing model among other planned models.

166
00:11:43,860 --> 00:11:46,950
So I have downloaded this video.

167
00:11:46,950 --> 00:11:47,940
So let me just.

168
00:11:49,430 --> 00:11:52,250
Let me just navigate my screen towards this video.

169
00:11:53,400 --> 00:11:58,470
So over here you can see that we are able to detect, uh, so this is a wrong detection truck.

170
00:11:58,770 --> 00:11:59,400
Uh, okay.

171
00:11:59,400 --> 00:12:01,620
And this is also a wrong detection person.

172
00:12:01,620 --> 00:12:02,880
This is not the person.

173
00:12:02,880 --> 00:12:05,040
And this is also a wrong detection.

174
00:12:05,040 --> 00:12:06,870
And this is not a truck as well.

175
00:12:06,870 --> 00:12:10,350
So this is other wrong detection okay.

176
00:12:10,350 --> 00:12:14,850
As we are able to detect the cars trucks as well.

177
00:12:14,850 --> 00:12:17,580
So the results look pretty fine to me.

178
00:12:17,910 --> 00:12:20,430
So let's go back as well.

179
00:12:20,430 --> 00:12:23,850
And now I will download other video.

180
00:12:24,210 --> 00:12:24,720
Uh okay.

181
00:12:24,720 --> 00:12:25,290
From here.

182
00:12:25,290 --> 00:12:30,990
So I will I will just downloaded the video from Google Drive into this Google Colab notebook and let's,

183
00:12:31,020 --> 00:12:37,440
uh, run uh, Glenn model on this input video, which is test Walker MP4.

184
00:12:37,440 --> 00:12:38,880
So let's run this.

185
00:12:44,230 --> 00:12:46,330
So I already told you, Glenn Stanford.

186
00:12:46,330 --> 00:12:49,210
Generalized efficient layer aggregation network.

187
00:12:49,960 --> 00:12:50,260
Okay.

188
00:12:52,440 --> 00:12:59,460
So the complete video is being divided into 341 frames, and we are doing detection on each of the frame

189
00:12:59,460 --> 00:13:01,020
one by one.

190
00:13:01,080 --> 00:13:03,930
So this might take few seconds to complete.

191
00:13:03,930 --> 00:13:06,660
And then we will display the output video over here.

192
00:13:09,510 --> 00:13:13,170
The inference on this dashboard that MP4 video is also done.

193
00:13:13,170 --> 00:13:16,410
And here is our output video.

194
00:13:16,410 --> 00:13:20,130
I have also displayed the output video into this Google Colab notebook.

195
00:13:20,130 --> 00:13:26,490
And let me download this video over here and let's see how our results look like.

196
00:13:26,490 --> 00:13:29,310
So let me just open this up okay.

197
00:13:29,310 --> 00:13:31,020
And let me navigate my screen.

198
00:13:31,020 --> 00:13:36,630
So over here you can see that we are able to detect the person's handbag over here.

199
00:13:36,630 --> 00:13:40,170
And um the results look pretty fine.

200
00:13:40,170 --> 00:13:43,320
Like you can see that uh, but there are some false positives as well.

201
00:13:43,320 --> 00:13:48,090
So if you just see over here, this is not a tennis racket or this is not a Frisbee.

202
00:13:48,090 --> 00:13:53,400
So there are some, uh, false positives as well, like you can see over here as well.

203
00:13:53,400 --> 00:13:57,270
It's detecting at tennis racket or Frisbee, which is wrong.

204
00:13:59,110 --> 00:13:59,740
Okay.

205
00:13:59,740 --> 00:14:01,720
And if we just go ahead.

206
00:14:01,720 --> 00:14:05,530
So now we will see how we can do inference on these two videos.

207
00:14:05,530 --> 00:14:12,820
Demo dot mp4 test dot mp4 using YOLO v9 extended model YOLO stands for you only look once.

208
00:14:13,330 --> 00:14:17,110
Okay, so if you just go over here in this GitHub repo.

209
00:14:17,110 --> 00:14:23,140
So if you want to do inference with the yolo v9 models, we need to write python dash dash dual.py.

210
00:14:23,200 --> 00:14:25,150
So we will run this.py file.

211
00:14:25,150 --> 00:14:29,350
And if you want to do inference with the clan models we just write detect.py.

212
00:14:29,350 --> 00:14:34,300
And if you want to do inference with the yolo v9 models we use detect dash dot py.

213
00:14:35,890 --> 00:14:42,370
So now I will just run this cell over here and I will do inference on this video.

214
00:14:42,610 --> 00:14:45,010
So let's see how does it goes.

215
00:14:46,790 --> 00:14:53,810
So the complete video is being divided into 1314 frames, and we will be doing detection on each of

216
00:14:53,810 --> 00:14:55,370
the frame one by one.

217
00:14:55,370 --> 00:15:02,750
I'm using, uh, the YOLO v9 extended model, the best or outperforming model among all the YOLO v9

218
00:15:02,750 --> 00:15:03,470
models.

219
00:15:04,980 --> 00:15:10,530
You can see that yolo v9 extended model, uh, gives mean average precision on the validation set of

220
00:15:10,530 --> 00:15:18,630
the data set is 55.6%, and it has more number of parameters as compared to the other YOLO v9 models

221
00:15:18,630 --> 00:15:21,300
like test 57.3 million parameters.

222
00:15:24,340 --> 00:15:26,890
So like it is about to complete.

223
00:15:26,890 --> 00:15:29,530
And let's see how does it goes.

224
00:15:30,490 --> 00:15:35,410
So, uh, 760 frames are done and let's wait for all the frames to complete.

225
00:15:35,410 --> 00:15:38,470
And I will also display the output video over here.

226
00:15:41,680 --> 00:15:47,530
For the inference on this demo video with Golovnin extended model is being completed, and I have displayed

227
00:15:47,530 --> 00:15:49,570
the output demo video over here as well.

228
00:15:49,570 --> 00:15:56,470
So let me just download it and let us compare these our these results with the clan model results as

229
00:15:56,470 --> 00:15:56,770
well.

230
00:15:56,770 --> 00:16:00,040
So I will do the comparison over here as well.

231
00:16:00,040 --> 00:16:02,560
So let me just navigate my screen.

232
00:16:02,560 --> 00:16:06,970
So over here you can see these are the results from the YOLO V9 extended model.

233
00:16:06,970 --> 00:16:08,350
This is our own detection.

234
00:16:08,560 --> 00:16:11,530
Uh the clan model is also detecting this as proc.

235
00:16:11,530 --> 00:16:13,360
But this is not the case over here.

236
00:16:13,360 --> 00:16:16,180
This is also a false positive okay.

237
00:16:16,390 --> 00:16:17,440
So.

238
00:16:18,580 --> 00:16:20,800
There are some false positives as well.

239
00:16:20,890 --> 00:16:26,770
Uh, but in the with the uh, Glen model, they are more false positives, I believe.

240
00:16:30,340 --> 00:16:34,960
So these are the results over here which you can see over here.

241
00:16:40,780 --> 00:16:44,440
So let's go back over here.

242
00:16:45,160 --> 00:16:52,210
Okay, so let's, uh, run on this other video test workload, MP4, and we will be using yolo v9 extended

243
00:16:52,210 --> 00:16:54,370
model again over here as well.

244
00:16:54,730 --> 00:16:59,650
So now I will just run this cell over here and let's see how our results look like.

245
00:17:02,010 --> 00:17:07,650
We on this video when we track an extended model, there were some false positive like it was detecting

246
00:17:07,650 --> 00:17:10,230
potholes as frisbees or tennis racket.

247
00:17:10,260 --> 00:17:13,890
Let's see if this is the case over here as well or not.

248
00:17:14,370 --> 00:17:17,340
So the complete video is divided into 341 frames.

249
00:17:17,340 --> 00:17:21,360
And detection on each of the frame is being done one by one.

250
00:17:21,360 --> 00:17:23,760
So let's see how does it goes.

251
00:17:32,100 --> 00:17:36,540
Now I will just display this output video into this Google Colab notebook as well.

252
00:17:36,540 --> 00:17:38,790
So this will take few more seconds.

253
00:17:39,330 --> 00:17:44,670
And here you can see all the models which I have downloaded from the repo into this Google Colab notebook.

254
00:17:44,670 --> 00:17:47,640
You will find all the names over here as well.

255
00:17:51,090 --> 00:17:54,330
So this might take a few more seconds before it gets completed.

256
00:18:09,360 --> 00:18:11,460
For the inference on this video is being done.

257
00:18:11,460 --> 00:18:12,990
And here we have the output video.

258
00:18:12,990 --> 00:18:16,050
So let me just download this up over here okay.

259
00:18:16,050 --> 00:18:22,470
So let me just navigate my screen towards this output video and let me just play it from start okay.

260
00:18:22,470 --> 00:18:26,130
So now you can see over here we don't have local.

261
00:18:26,130 --> 00:18:28,080
They are very less false positives okay.

262
00:18:28,080 --> 00:18:30,120
There is one false positive over here.

263
00:18:30,390 --> 00:18:33,030
Else overall the results look pretty fine.

264
00:18:33,030 --> 00:18:38,430
Like in the case of Glen extended model there were more number of false positives like here.

265
00:18:38,430 --> 00:18:41,430
And in this case we have quite less false positives.

266
00:18:41,430 --> 00:18:42,930
So that's so good.

267
00:18:42,930 --> 00:18:48,840
So in this tutorial we have seen that how we can do object detection using Glen and YOLO benign models.

268
00:18:48,840 --> 00:18:55,500
And we have seen that how we can do uh run these models in Google Colab and do object detection on images

269
00:18:55,500 --> 00:18:56,460
and videos.

270
00:18:56,460 --> 00:18:58,290
So that's all from this tutorial.

271
00:18:58,290 --> 00:18:59,400
Thank you for watching.