Official YOLO v7 Object Detection COMPLETE Tutorial for Google Colab

This YOLO v7 tutorial enables you to run object detection in colab. This is a complete tutorial and covers all variations of the YOLO v7 object detector. YOLO v7 has just been released and shows accuracy and speed improvements over its predecessors. Currently, YOLO v7 is the world's fastest and most accurate object detector. This tutorial is focused on running pre-trained YOLO v7 models (all seven variants).
YOLO v7 surpasses all known object detectors in both speed and accuracy in the range from 5 FPS to 160 FPS and has the highest accuracy 56.8% AP among all known real-time object detectors with 30 FPS or higher on GPU V100.
* Code is available for our Patreon Supporters*
/ thecodingbug
---------------------------------------------
► Time Stamps:
Introduction: (0:00)
Setting up Colab Notebook: (0:22)
Mounting Google Drive: (0:44)
Create YOLOv7 Directories: (1:37)
Download YOLOv7 Model: (2:46)
Modify Code Files: (3:21)
Copy Image and Video: (6:20)
YOLOv7 Object Detection on Images: (6:33)
YOLOv7 Object Detection on Videos: (7:15)
---------------------------------------------
► Links:
github.com/WongKinYiu/yolov7
---------------------------------------------
Want to discuss more?
►Join my discord: / discord
#TheCodingBug
---------------------------------------------
► My Other Tutorials:
○ YOLOv7 Custom Object Detection (Win, Linux): • Official YOLO v7 Custo...
○ YOLOv7 Complete Tutorial (Windows, Linux): • Official YOLO v7 COMPL...
○ 40 Object Detection Models in TensorFlow: • TensorFlow Object Dete...
○ Realtime Object Detection on CPU with OpenCV: • Realtime Object Detect...
○ DETECTRON2 Custom Object Detection, Custom Instance Segmentation Tutorial: Part 2 (Training Models): • DETECTRON2 Custom Obje...
○ DETECTRON2 Custom Object Detection, Custom Instance Segmentation Tutorial: Part 1 (Preparing Data): • DETECTRON2 Custom Obje...
○ Detectron2 on Colab: • DETECTRON2 TUTORIAL fo...
○ Instance Segmentation as Rendering: • DETECTRON2 PointRend T...
○ Detectron2 Complete Tutorial: • COMPLETE DETECTRON2 TU...
○ YOLOv4 On Android Using TFLite: • YOLOv4 TFLite Object D...
○ Custom YOLOv4 Object Detection with TensorFlow and TFLite : • Custom Object Detectio...
○ Darknet YOLOv4 Custom Object Detection: Part 2 (Training YOLOv4 Darknet): • YOLOv4 Custom Object D...
○ Darknet YOLOv4 Custom Object Detection: Part 1 (Preparing Custom Dataset): • YOLOv4 Custom Object D...
○ YOLOv4 Object Detection with TensorFlow, TFLite and TensorRT: • YOLOv4 Object Detectio...
○ Darknet YOLOv4 Object Detection for Windows 10 on Images, Videos, and Webcams: • Darknet YOLOv4 Object ...
○ Real-Time Object Detection on Webcam and Videos Using OpenCV With YOLOv3 and YOLOv4 | Windows Linux: • YOLOv4 and YOLOv3 Obje...
○ Real-time Multiple Object Tracking with YOLOv4 TensorFlow and Deep Sort | Linux, Windows: • Object Tracking with T...
---------------------------------------------
► Follow us on Twitter: / bugcodingthe
► Support us on Patreon: / thecodingbug
---------------------------------------------
DISCLAIMER: Links included in this description might be affiliate links. If you purchase a product or service with the links that I provide I may receive a small commission. There is no additional charge to you!

[@junaidqadirb]

You can also double-click a file to edit it within Google colab and save it.

[@TheCodingBug]

Pinning the comment as it'll be helpful for others.

[@martinkwesimawusi8370]

I have been looking for the best video to properly introduce me to google colab and YOLO and you have just done that to my sactisfaction. Thank you very much.

[@iamxba3a817]

okay so this is literally one tutorial in whole yt that helped me get that video object tracking.. Im so thankful !!!!!

[@zongzong0408]

That is useful, thanks! Showing FPS is important for my work.

[@hugehammer2706]

Damm man! Nice video, I'm finally starting object detection

[@182Kandy]

good afternoon! thank you for making this content available, it helped a lot. Do you happen to know if it would be possible to detect tree species in a native forest, with a drone orthomosaic?

[@amirtarek7166]

I have a problem while training the model, the epoch suddenly stop while extracting certain batches and take around 25 mins to continue , what could lead to such a problem ? and how can i over come it ?

[@code786yt2]

Thank you very much Sir. I appericiate you.

[@LimChinHong-ju6fo]

good video sir, may I know that how to train the model on a batch of images and evaluate them all at once using things like mAP? thank you sir.

[@esooghazy]

Fast and useful! Thanks a lot!

[@TheCodingBug]

Thank you for the appreciation.

[@sabrimas9718]

can you make a video on how to modify yaml architecture in yolov8? i'm doing research and want to add attention modules like GAM, CBAM and others but don't know where to start and when i change it, the called architecture configuration stays the same, and there is no change.

[@ajayreddykanduru453]

Hi, can you please share how to take input from the webcam ?

[@tristanrajendraa]

can yolo use for image classification and optimezer using adam,sgd,rmsprop?? my dataset is rice so i want yolo to detect it like it's basmathi rice or regular rice

[@azimbekkhudoyberdiev0817]

Thank you for best resources. How we can get mAP results for object detection in videos and collect them as a file?

[@jaiborade3421]

Hi sir, thanks for the great session. If possible please share custom dataset training

[@user-tu3ch9xe4q]

any method to get the coordinates of boxes formed

[@muhammadmudasir1166]

May God bless you sir thank you so much ❤❤

[@phunghm]

You're my best youtube teacher! If possible, tell how to illiterate the output folder (not exp) and this folder is initialized as a variable in run time. It works when I change save_dir to a fix path in detect.py but couldn't do if assign a changeable value to save_dir. Thank you!

[@EdenW0311]

Thank you very much

[@deepasavant821]

time taken to detect an object is in seconds?

[@MinhazulAbedinMunna]

Here, Can we count the object from image ? it will be helpful

[@tranquynhat4933]

When I add photos to yolov7, I run the command !python detect.py --weight yolov7.pt --conf 0.5 --img-size 640 --source m_caysi.jpg gives this error python3: can't open file '/content/detect.py': [Errno 2] No such file or directory . Do you have any way to fix it? Thank you. Fastest response please help me

[@pvrohanraj]

Great Video, Simple explanation. No errors. why did you comment the datasets print? was there a specific reason?

[@TheCodingBug]

Which print are you referring to? Can you point out the time stamp?

[@barbaralau3254]

Hi, Thank you for your video. I followed the steps. I got the test_batch0_labels.jpg in the last exp folder inside runs/detect folder. I can see the labels and boxes correctly surrounds the object in this image. Does it mean the training is successful or this is not the right place to look into? My question is I ran the detect.py command with the test image and best.pt , I didn't get back any boxes or labels at all. Can you please tell me why does it happen? If the test_batch0_labels.jpg exists, shall I expect this training is completed? or how to fix this problem? Thank you for your help!

[@TheCodingBug]

Check precision recall and f1 curves. If those have a very small value, model learned nothing and you need more data to train.

[@AmitKumar-hm4gx]

I have a question, When we run inference is there any specific reason authors have decided to print YOLOR model ? is Yolo7 same as YOLOR or if there is similarity in terms of certain architectures used ?

[@TheCodingBug]

Thats because the implementation of YOLOv7 is based on YOLOR and they use many things from that code. Perhaps they missed this print statement and did not change it.

[@AmitKumar-hm4gx]

@@TheCodingBug thank you :)

[@abrarluvrabit]

if we want to show FPS on yolov5 and yolov6 do we need to follow the same thing as showed for yolov7 or not? if not then please make a video for FPS showing in yolov5 and 6 also it will be very helpfull. thanks

[@TheCodingBug]

We will do the same thing.

[@rizvee3977]

'git' is not recognized as an internal or external command, operable program or batch file. Traceback (most recent call last): File "C:\Users\User\yolov7\utils\google_utils.py", line 26, in attempt_download assets = [x['name'] for x in response['assets']] # release assets KeyError: 'assets I am getting this error. How to solve this problem?

[@TheCodingBug]

Download and install git for your operating system.

[@karlomarkovic2404]

Hi, thanks for the great tutorial. How could you measure the accuracy and get the metrics like mean average precision? Could you do a tutorial on that?

[@TheCodingBug]

Those would already be present in training folder (where best weights are) along with precision recall etc graphs.

[@karlomarkovic2404]

@@TheCodingBug I appreciate the answer, but I am still a newbie to this. I would really appreciate if you could give ma more detailed explanation. I would like to try something like this for my final thesis, but I am not sure how to objectively measure the accuracy. THanks!

[@vic0th380]

@@karlomarkovic2404 You split the dataset and train on some data (train dataset) and test your model on remaining data. Then you compare the results with test dataset, calculate how accurate your results are. You also need to avoid model overfitting which might result in high accuracy by working quite well on the test dataset but would be problematic with other datasets.

[@abrarluvrabit]

@@vic0th380 thanks brother

[@anaghasm7584]

It would be really helpful if you could teach how to detect objects on Webcam using google colab.

[@abrarluvrabit]

hello, did you got any video regarding to your question please share if you got it. thanks

[@its_mjleo]

How can we run the v7 detection on webcam in Google Colab?

[@abrarluvrabit]

did you got any solution? i also need to try that please share if you know.

[@lpmuinsisamarinda]

thks for tutor

[@ikramessafi9560]

Thank you a lot for that video, but i want to know if o have an other object that i wanna detect like fruits how can i use this model to detect it?

[@TheCodingBug]

You'll need to train custom model on custom data: kzbin.info/www/bejne/Y4K6qX1mn554j5o

[@ikramessafi9560]

@@TheCodingBug Thank you a lot , i wanna also ask you if you have any idea about counting those objects that are detected .?

[@maheshKumar-iw4qb]

Real time detection how

[@danialabyazi6697]

Thank you for the great introduction of YOLOV7 code. Do you know how we can access to detected boundary box information from code?

[@abrarluvrabit]

Hi daniyal did you get the ans of your question? because i also want to know the same thing so if you know then plz share it with me

[@xniux]

--save_txt you will see (I guess) one line for each object in "class, x, y, width, height" in one txt file for a frame if object(s) detected

[@derwinmahardika3754]

Can we use another version model such as yolov3 etc?

[@TheCodingBug]

No not with this code.

[@freescholerships_2.0]

Hi, We need dataset for object Detection using YOLOv7

[@kurumallasatyanarayana9226]

Yes

[@kurumallasatyanarayana9226]

Please tell me the dataset. Iss mscoco we can download in Google drive from kaggle

[@muhammadmudasir1166]

Will it work on any random image

[@TheCodingBug]

Yes

[@robertraafat413]

Bro how can i convert .yaml and .pt to .cfg and .weights?

[@TheCodingBug]

Right click, rename, change extensions 😁 Just kidding. I understand that you want to convert it to darknet yolo format. The cfg and weights files for pretrained model is available with the official repository. For custom model, you'll need to build darknet and then train the model from scratch. See if you can get it done with existing yolov4 darknet tutorial.

[@robertraafat413]

@@TheCodingBug hmm so how to build it, can you make a video on that

[@fahmidanial]

Why I got attribute error when running on street2.mp4 "module cv2 has no attribute font_hersyey......"

[@fahmidanial]

Please help

[@TheCodingBug]

Your spelling is wrong.

[@fahmidanial]

@@TheCodingBug hahaha thank you so much for your respond

[@abrarluvrabit]

can you please share this video file street2.mp4 with me.

[@bcheng32]

Can you share how to find 2.jpg and street2.mp4? Please help if you see it, thank you sir

[@omoklamok]

hi there its good to see yolo7 new subscriber here... i tried to follow but somehow i got some error

[@TheCodingBug]

Hi. What's the error?

[@omoklamok]

wow you replied. wait sir i will re run it again

[@omoklamok]

@@TheCodingBug File "D:\YOLOV7\detect.py", line 196, in detect() File "D:\YOLOV7\detect.py", line 26, in detect (save_dir / 'labels' if save_txt else save_dir).mkdir(parents=True, exist_ok=True) # make dir File "C:\Users\MSC\.conda\envs\yolov7\lib\pathlib.py", line 1327, in mkdir self.parent.mkdir(parents=True, exist_ok=True) File "C:\Users\MSC\.conda\envs\yolov7\lib\pathlib.py", line 1328, in mkdir self.mkdir(mode, parents=False, exist_ok=exist_ok) File "C:\Users\MSC\.conda\envs\yolov7\lib\pathlib.py", line 1323, in mkdir self._accessor.mkdir(self, mode) FileNotFoundError: [WinError 2] The system cannot find the file specified: 'runs\\detect'

[@k.v.srilakshmiasharani1553]

what is 2.jpeg ?

[@TheCodingBug]

It's an image to run object detection on.

[@kurumallasatyanarayana9226]

Is any imagr can be added to drive for detection??

[@kurumallasatyanarayana9226]

Any images can be added in drive and shall we detect it???

[@TheCodingBug]

@@kurumallasatyanarayana9226 Yes

[@kurumallasatyanarayana9226]

@@TheCodingBug here which dataset used sir