Did you enjoy this video? Try my premium courses! 😃🙌😊 ● Hands-On Computer Vision in the Cloud: Building an AWS-based Real Time Number Plate Recognition System bit.ly/3RXrE1Y ● End-To-End Computer Vision: Build and Deploy a Video Summarization API bit.ly/3tyQX0M ● Computer Vision on Edge: Real Time Number Plate Recognition on an Edge Device bit.ly/4dYodA7 ● Machine Learning Entrepreneur: How to start your entrepreneurial journey as a freelancer and content creator bit.ly/4bFLeaC Learn to create AI-based prototypes in the Computer Vision School! www.computervision.school 😃🚀🎓
@@ComputerVisionEngineer hi Felipe, will this wonderful tutorial work with the current YOLOv8 version and its dependencies? or we need the exact versions you used in this video? will there be errors? compatibility issues?
The video is of high quality, simple and practical, and avoids a lot of boring and difficult computer knowledge. As a beginner, I feel that I have received new knowledge, thank you.
the fact that you have shown how to do it in both colab and locally is just amazing. you are sooo underrated. please don't stop making content. your content is sooo awesome
I used another random dataset with 1000 images in train and 100 images in validation, and I tried to keep val and train datasets different in the config.yaml. The results generated on every graph in (results.png) consist of a single dot. Should I choose the train dataset and validation dataset for correct results? Can you please help me? Regards🌼🌻🌼 *Edit: I got my issue resolved; I just used 20 epochs instead of one Thanks for such a great and only video on yolov8 that is crystal clear about every point.
I did everything like on video but with different images and path. Checked everything multiple times but when i run my main python file nothing happends😢
hello, when i run 'create_dataset_yolo_format.py' I get an error saying "The system cannot find the path specified: '.\\data\\train'". Could u pls help me with this? Thanks
thanks Felipe, as a noob this really helpful... something that i wanna know, how much picture that we need in yolov8n? how much picture of alpaca that u have? are we just submit 14 picture to the cvat or you must submit all picture?
Great Video for beginners ♥ A video about training an actual useful dataset like escooters and escooters + driver would be cool. If not possible with training, do it with code, like checking if a person and escooter box intersect, etc. And detecting multiple people standing on one scooter and stuff like that. I mean i love alpacas, but i cannot think about a usecase for that or some video to build on top of the previous ones. Just a suggestion to make it more interesting as real projects could be a great YT series :) Greetings from Vienna
@@ComputerVisionEngineer thanks for your answer, i hope you did not misunderstood it, you have great in-depth videos of course and i love them. I just meant to give beginners a more guided journey in form of a series with more practical use case/s. It might still sounds weird and like critique, but it's really just an idea with my chaotic english skills :D Greetings
thank you for the amazing tutorial! I was able to train YoloV8 on a custom data, but I was wondering how I can continue to train it using weights from previous training sessions. In other words, how do i make the second training session include the knowledge it gained from the first training session?
Howdy! I have a question. When training a model, are we using our own hardware, or is it like using Google Colab (I'm guessing that when using Colab we don't use our own hardware)?
Thanks for this comprehensive tutorial. What if the goal is to add an additional class to the available classes in YOLOV8. Particularly, annotation tool starts annotating from label number 0. However, class 0 is available in corresponding classes in YOLOV8. does CVAT allow to label from the last index of COCO dataset to add a new class to the available ones in YOLOV8?
i have a question if i am trainig a model for license plate detection and it is taking 8 hours trainig for 1 epoch tell me how you trained your model for the license plate detection in the other video i am stuck at this point i have to submit my project at the end of this week please help me with this
Resumen del video [00:00:00][^1^][1] - [00:37:03][^2^][2]: Este video proporciona una guía detallada sobre cómo entrenar un detector de objetos Yolov8 en un conjunto de datos personalizado. El proceso incluye la recolección y anotación de datos, estructuración de datos en el formato requerido por Yolov8, y finalmente, el entrenamiento del modelo. **Destacados**: + [00:00:00][^3^][3] **Introducción al entrenamiento de Yolov8** * Presentación del proceso completo de entrenamiento * Importancia de la recolección de datos adecuados + [00:05:06][^4^][4] **Anotación de datos** * Uso de la herramienta de anotación CVAT * Proceso de etiquetado de imágenes con cajas delimitadoras + [00:19:54][^5^][5] **Estructuración de datos para Yolov8** * Formato específico requerido para el entrenamiento * Creación de directorios de imágenes y etiquetas + [00:30:03][^6^][6] **Entrenamiento del modelo Yolov8** * Uso del repositorio oficial de Yolov8 * Métodos para entrenar el modelo en un entorno local o en Google Colab + [00:37:12][^3^][3] **Configuración inicial** * Establecimiento del directorio raíz y especificación de datos de entrenamiento y validación * Uso de los mismos datos para simplificar el tutorial + [00:38:10][^4^][4] **Entrenamiento en Python** * Ejecución del entrenamiento con Yolov8 para un solo epoch como demostración * Observación del proceso de carga de datos y entrenamiento + [00:43:03][^5^][5] **Entrenamiento en Google Colab** * Creación de un cuaderno en Google Colab y montaje de Google Drive * Instalación de la biblioteca ultralytics y configuración del entrenamiento + [00:50:06][^6^][6] **Evaluación del modelo** * Revisión de los resultados del entrenamiento y análisis de la pérdida * Pruebas prácticas con videos de alpacas para evaluar la precisión del modelo
Awesome video, thanks for talking the time to make it! Would you be able to suggest an interesting way to extend the YOLO algorithm? I am researching ideas for a Masters project in AI…
Thank you for sharing. I have question about results of YOLOv8 model , after the training of the model it results in a 3-dimensional confusion matrix taking the background as a class knowing that I have a binary classification my project is classification of preforms whether it is defect or not. What can be the raisaon of appearance of this class "backgroud" and how I can solve it ? if you help me I would be grateful.
I see this is listed as an issue in ultralytics github repository. No answer was given. I think you would need to dig into the code to see what is going on. 🙌
i had problem to install the python lib called ultralytics. how i can install it on my system. i type in cmd pip install ultralytics. it showing the the msg called successfully install but in vs code it showing error
I want to ask something. How can you training your model without making variabel to your relative annotation path and just training yolov8n with your relative images path? Oh ya and i just remember. Why your relative path to image dataset is Images/train and not /Images/train?
Hi, not sure if I understand your question. The config file has both variable, the absolute path to the root directory and the relative path to the images directory. 🙌
What happen if a fine tune a model with class 0 alpaca, then yolo will not recognize none of the others classes? person that is the actual 0 will be change by 0 alpaca? What if i want to add another class how would I do it?
after training the model when i ran the last cell i got this error NotImplementedError: A UTF-8 locale is required. Got ANSI_X3.4-1968 but it was fixed by running the following code before just before the last cell import locale locale.getpreferredencoding = lambda: "UTF-8"
Hello! I'm having some trouble doing the configuration of the dataset. My code for the yaml file is here below: path: C:\Ze\\2024\PosIFSC\TCC\Python\Code\dataset train: train\images val: valid\images names: 0: pig That is the error presented on the prompt. RuntimeError: Dataset 'C://Ze/2024/PosIFSC/TCC/Python/Code/data.yaml' error object of type 'NoneType' has no len() I'm using VSCode on Windows 11. I've tried everything but I just can't get it right. Please help me!
hi, I have a question. if I divide the dataset into training and testing only, is it necessary to run the validation part? and if not, during inference how to find out the mAP? plis help me🙏
TY very much for the great video. Edge case question: I'm training my model on only 10 images and I'm never getting any predictions even tho my images are well labeled. Do you have an idea why?
Extremely useful tutorial for which I thank you. Question please : I looked at your Gti account and could not find the script you are talking about at 1:00:15 ? Thank you. Philippe
@Computer vision engineer Will this work for YOLOv5, working on a university projekt where I use a Stereo Vision camera to detect boats and measure the distance and classify them, this video is perfekt for my case. Can I use it in YOLOv5?
thanks for the video ! however I followed all your steps and eventually got nothing in the pic with predictions. no rectangle. I'm wondering if it has to do with the number of photos I trained. only used 6... if you have any advice on where the problem could come from, it would be super nice !
I just want to make sure that I just need the main file and the config file that you have presented, right ? other files (predict.py, results.py...) are not requested, aren't they ? because I also saw other lines of code in your link below so I just don't want to be confused... thanks again@@ComputerVisionEngineer
can i know the source from which u downloaded the image dataset other than the one u mentioned in the video cause i want to train the model to detect Ash trees ,any suggestions would be of great help thank u
Hi, if you are referring to an edge device, there are multiple ways to run it. In one of my recent videos I show you how to run it in a Raspberry Pi and a coral accelerator. 🙌
@@ComputerVisionEngineer Oh thx. So some "tops" should be enough in this context. I'm trying to get the correct hardware to train but it's not clear to me to run it on production, maybe I was thinking of the rockchip with 6 tops I'll start from there and see in practice what hardware I need.
Hi.. I have a dataset consisting of png images(no bounding boxes) and annotations of corresponding images as. xml files(PASCAL VOC).. HOW SHOULD I TRAIN? ALL TUTORIALS I SEE ARE OF JPG AND TXT.. PLEASE HELP
Im new, and I always get this problem said 'FileNotFoundError: [WinError 2] The system cannot find the file specified' When i run it on my VSCode. I've made the config.yml file, anyone have any suggestion? please and thank you
take a look at the script predict_video.py, also take a look at my other tutorial on object detection + tracking with deep sort, I make inferences with unseen images in that video. 🙌
Hi mr. I want to automate palm tree counting using yolo v8. Can the result of the testing export as shp file so i can generate the coordinate from all the detected tree?
i really like your tutorials , but i have an issue during the custom object training i tried it with google colab and also with pycharm but in the pred. image it doesn't show the prediction neither rectangle box nor name tag. please help bro .................
Instead of rectangular bounding box I need instance segmentation, can you make video on instance segmentation with keypoints for custom datasets, please
Why do I keep getting this error. RuntimeError: Dataset 'config_2.yaml' error config_2.yaml 'train:' key missing . 'train' and 'val' are required in all data YAMLs. I use a windows computer. Is there a unique way of specifying the file path in the yaml file?
Hi any idea how this task can be done :When working on annotating lines that connect the light sources , i used polyline tool of cvat , annotated all the pairing correctly , downloaded it in yolo format but the text files are empty, which was not the case while drawing boxes around light sources.Any idea where i might be going wrong or how to save annotaions for polylines to detect the lines in image. So how should this task can be done? I searched many ways but couldn't find any way please look into it . Thank you.
Hey, for polylines you may need to download your annotations in a different format, for example in cvat format. I think it will be very challenging to detect lines using an object detector like yolo. It sounds like a very niche use case, take a look at papers and other resources online, from the top of my head I cannot think in any way to detect lines using an object detector.
@@ComputerVisionEngineer Thank you for your response , So any idea like how can I detect lines joining light sources in a floor plan as using yolo will not work as I cannot get the annotation coordinates?any other method you think that should work for this task or any model that accepts cvat format?
The video is good and simple EXCEPT FOR the fact that you used the same data for train and validation. Don't do that! That is arguably missleading. It wouldn't be that hard to teach your audiacne what the point of train, test and val is and using the same data for train and val is veerry missleading and confusing since it dosn't really give acurate results.
Thank you so much for your feedback. The only reason I used the same data for training and validation was for simplicity, because I wanted to focus on the overall process of training an object detector. I see your point. I will definitely take it into account for future videos. 🙌
HELLO SIR!!! GREAT TUTORIAL. I am creating an android application, to detect the bounding boxes and getting the coordinates, how can i do so, can you help me out!!
Thank you very much for a great tutorial! Great explained and easy to follow. I've now trained a custom model (in Google Colab) for 3 classes. When testing it with unseen pictures, I get good results. If I were to improve my custom model, how do I do that in Google Colab, change the "model = " part and point it to my custom .yaml file? What should I do with my pictures in the train and val folders, keep the old ones that I trained the model on and add the new ones, or empty them and only add new pictures? Should I also add new pictures to the val folder?
I am glad the video was useful! 😃 Keep the old pictures you trained the model on and add new images, yes it is convenient to add images in the val directory too. 80/20 is a good ratio of images in train/val, so if you add images in the train dir also add pictures in val so you maintain the ratio. 🙌
@@ComputerVisionEngineer Thanks! Worked like a charm! Another question, if you happen to read it 🙂 I'm thinking of using YOLOv8 to look for defects on an object. How should I go forward with labeling and which classes to use? Should I use one class to put a label around the whole object without defects -> class "OK_object"? What about when it is defect(s) on the object, should I then label only the defect(s) -> class "Defect" and not label the whole object in addition? Or should I just label defect(s) and not label obejct(s) where no defect(s) is found, so that objects without defect(s) will be 'background' when training my model?
Hi Felipe , I have run into problem, since I successfully compiled my code but unfortunately it wasn't showing the *runs* folder which leaded to the results, plz help me out.
@@ComputerVisionEngineer Yup I fixed it, All that time I saved the Folder *Label* 'L' as starting letter as capital instead of small letter 'l' which I should have done. But I corrected it and now I am gonna start working on the License reader from one of your videos on the same 👍
Heyaaa! It's an awesome tutorial! I have a doubt though... now I am getting the output MAP50 for all the classes as a whole. But in my project, there are two classes, how to get the individual mAP50 for each of them? Thanks in advance:)
Hey, if it is not provided in yolov8 results you may need to compute it yourself. Take a look at this repo github.com/Cartucho/mAP.git, it seems exactly what you need. 💪💪
Did you enjoy this video? Try my premium courses! 😃🙌😊
● Hands-On Computer Vision in the Cloud: Building an AWS-based Real Time Number Plate Recognition System bit.ly/3RXrE1Y
● End-To-End Computer Vision: Build and Deploy a Video Summarization API bit.ly/3tyQX0M
● Computer Vision on Edge: Real Time Number Plate Recognition on an Edge Device bit.ly/4dYodA7
● Machine Learning Entrepreneur: How to start your entrepreneurial journey as a freelancer and content creator bit.ly/4bFLeaC
Learn to create AI-based prototypes in the Computer Vision School! www.computervision.school 😃🚀🎓
did u aimbot the alpaca ?
@@johnmcook1 😂 I did not
@@ComputerVisionEngineer hi Felipe, will this wonderful tutorial work with the current YOLOv8 version and its dependencies? or we need the exact versions you used in this video? will there be errors? compatibility issues?
,
Hi...can you please show details how to download and create dataset from openimagedatasetV7
The video is of high quality, simple and practical, and avoids a lot of boring and difficult computer knowledge. As a beginner, I feel that I have received new knowledge, thank you.
Thank you!! So glad you enjoyed it and that you found the content helpful! 😃🙌
youre so fluid coherent and elaborate 1 hour felt like 1 minute
😊 Thank you so much for your kind words! Glad you enjoyed the video! 🙌
the fact that you have shown how to do it in both colab and locally is just amazing. you are sooo underrated. please don't stop making content. your content is sooo awesome
Your videos are really simple and shares good amount of knowledge, as a beginner your videos always helped me a lot
I really like watching your tutorials, you have great energy 🥳 and a good way of teaching the concepts 🤓
😃💪💪💪
Great video, very detailed. Really stands out from a lot of very confusing and crappy tutorials. Also your accent is so interesting.
I sincerely appreciate your indepth turial. It kicked started my journery in Computer Vision.
Hi, nice tutorial. I really enjoyed it. Is it possible for you to make a multi-label classification tutorial using DL algorithms?
Hey, thank you, I am happy you enjoyed it! 😊 Sure, that is a great suggestion, I will try to do a video about that soon! 😃💪
Spectacular tutorial, sir. Thank you.
Hi Felipe, Ultralytics using any standardscale for detecting object.
Original images to standard format
I used another random dataset with 1000 images in train and 100 images in validation, and I tried to keep val and train datasets different in the config.yaml. The results generated on every graph in (results.png) consist of a single dot. Should I choose the train dataset and validation dataset for correct results? Can you please help me?
Regards🌼🌻🌼
*Edit: I got my issue resolved; I just used 20 epochs instead of one
Thanks for such a great and only video on yolov8 that is crystal clear about every point.
I did everything like on video but with different images and path. Checked everything multiple times but when i run my main python file nothing happends😢
Thanks a lot!! big thanks for teaching how to do it on a local ide
Superb tutorial! The best I've seen yet.
This is fantastic! Excited to give this a shot-appreciate it, mate!
😃 You are welcome!
@@ComputerVisionEngineer do you have any good resources or suggestions for taking this trained model and then doing object counting in an ROI?
hello, when i run 'create_dataset_yolo_format.py' I get an error saying "The system cannot find the path specified: '.\\data\\train'". Could u pls help me with this? Thanks
thanks Felipe, as a noob this really helpful... something that i wanna know, how much picture that we need in yolov8n? how much picture of alpaca that u have? are we just submit 14 picture to the cvat or you must submit all picture?
TYSMMM this absolutely helped me create my first object detection model TYSMMMMMMM
You are welcome! 😃 Glad you find it useful! 🙌
@@ComputerVisionEngineer yep defiantly the best course abt the topic on youtube
@@ComputerVisionEngineer are you considering doing a yolo nas course??
Abosolutely helpful! So understandable and thank you so much!!!
You are welcome! Glad it is useful! 😃🙌
Great Video for beginners ♥ A video about training an actual useful dataset like escooters and escooters + driver would be cool. If not possible with training, do it with code, like checking if a person and escooter box intersect, etc. And detecting multiple people standing on one scooter and stuff like that. I mean i love alpacas, but i cannot think about a usecase for that or some video to build on top of the previous ones. Just a suggestion to make it more interesting as real projects could be a great YT series :) Greetings from Vienna
Thank you for your recommendations! I will try to make a more practical use case of object detection. Cheers! 😃🙌
@@ComputerVisionEngineer thanks for your answer, i hope you did not misunderstood it, you have great in-depth videos of course and i love them. I just meant to give beginners a more guided journey in form of a series with more practical use case/s. It might still sounds weird and like critique, but it's really just an idea with my chaotic english skills :D Greetings
Thank you for all your efforts, i wanna ask you can you give us a roadmap to learn machine learning ?
Thank you another time
Hey, sure, I will be happy to make a machine learning roadmap! 😃💪
You are many many many (cit.) the best! 😂😂! Thank you!
😂 Thank you for your support! 😃🙌
thank you for the amazing tutorial! I was able to train YoloV8 on a custom data, but I was wondering how I can continue to train it using weights from previous training sessions. In other words, how do i make the second training session include the knowledge it gained from the first training session?
how show results using webcam ?
I get error with :
while True:
ret, frame = cap.read()
result = model(frame)
cv2.imshow("yolov8", frame)
Great indepth explanation of the entire process. 👍👍
Glad it was helpful!! 😃💪
Hi, I have a question, I want to use a picture instead of using a video for the prediction stage, is that possible? And If it is, how should I do it?
Howdy! I have a question. When training a model, are we using our own hardware, or is it like using Google Colab (I'm guessing that when using Colab we don't use our own hardware)?
Thanks for this comprehensive tutorial. What if the goal is to add an additional class to the available classes in YOLOV8. Particularly, annotation tool starts annotating from label number 0. However, class 0 is available in corresponding classes in YOLOV8. does CVAT allow to label from the last index of COCO dataset to add a new class to the available ones in YOLOV8?
i have a question if i am trainig a model for license plate detection and it is taking 8 hours trainig for 1 epoch tell me how you trained your model for the license plate detection in the other video i am stuck at this point i have to submit my project at the end of this week please help me with this
I have used the same images but it is not detecting the alpaca in the output video
Great Job... Keep Posting...
hello, do you have a video for running yolo v9 locally?
Resumen del video [00:00:00][^1^][1] - [00:37:03][^2^][2]:
Este video proporciona una guía detallada sobre cómo entrenar un detector de objetos Yolov8 en un conjunto de datos personalizado. El proceso incluye la recolección y anotación de datos, estructuración de datos en el formato requerido por Yolov8, y finalmente, el entrenamiento del modelo.
**Destacados**:
+ [00:00:00][^3^][3] **Introducción al entrenamiento de Yolov8**
* Presentación del proceso completo de entrenamiento
* Importancia de la recolección de datos adecuados
+ [00:05:06][^4^][4] **Anotación de datos**
* Uso de la herramienta de anotación CVAT
* Proceso de etiquetado de imágenes con cajas delimitadoras
+ [00:19:54][^5^][5] **Estructuración de datos para Yolov8**
* Formato específico requerido para el entrenamiento
* Creación de directorios de imágenes y etiquetas
+ [00:30:03][^6^][6] **Entrenamiento del modelo Yolov8**
* Uso del repositorio oficial de Yolov8
* Métodos para entrenar el modelo en un entorno local o en Google Colab
+ [00:37:12][^3^][3] **Configuración inicial**
* Establecimiento del directorio raíz y especificación de datos de entrenamiento y validación
* Uso de los mismos datos para simplificar el tutorial
+ [00:38:10][^4^][4] **Entrenamiento en Python**
* Ejecución del entrenamiento con Yolov8 para un solo epoch como demostración
* Observación del proceso de carga de datos y entrenamiento
+ [00:43:03][^5^][5] **Entrenamiento en Google Colab**
* Creación de un cuaderno en Google Colab y montaje de Google Drive
* Instalación de la biblioteca ultralytics y configuración del entrenamiento
+ [00:50:06][^6^][6] **Evaluación del modelo**
* Revisión de los resultados del entrenamiento y análisis de la pérdida
* Pruebas prácticas con videos de alpacas para evaluar la precisión del modelo
Awesome video, thanks for talking the time to make it! Would you be able to suggest an interesting way to extend the YOLO algorithm? I am researching ideas for a Masters project in AI…
Thank you! 😃 Not sure if I would be able to suggest ideas to extend the YOLO algorithm, but best of luck in your masters degree! 🙌
Thank you for sharing. I have question about results of YOLOv8 model , after the training of the model it results in a 3-dimensional confusion matrix taking the background as a class knowing that I have a binary classification my project is classification of preforms whether it is defect or not. What can be the raisaon of appearance of this class "backgroud" and how I can solve it ? if you help me I would be grateful.
I see this is listed as an issue in ultralytics github repository. No answer was given. I think you would need to dig into the code to see what is going on. 🙌
i had problem to install the python lib called ultralytics. how i can install it on my system. i type in cmd pip install ultralytics. it showing the the msg called successfully install but in vs code it showing error
I want to ask something. How can you training your model without making variabel to your relative annotation path and just training yolov8n with your relative images path?
Oh ya and i just remember. Why your relative path to image dataset is Images/train and not /Images/train?
Hi, not sure if I understand your question. The config file has both variable, the absolute path to the root directory and the relative path to the images directory. 🙌
This also allow us to do the live detection?
This tutorial was very helpful for me. Thank you! Is it possible to learn more about visual relationships?
😊 I am glad you found it helpful! Training a model with visual relationships type of annotations would be very cool! I will try to do it. 🙌
nice mastery of the english language
What happen if a fine tune a model with class 0 alpaca, then yolo will not recognize none of the others classes? person that is the actual 0 will be change by 0 alpaca? What if i want to add another class how would I do it?
Amazing tutorial and explication, +1 subscriber
Thanks For the tutorial, But how to downloaded all the images into out local file, like the alpacas in your local device
Can I use 3d models for education? and than work with photo recognition?
I mean for multiple objects much simple find 3d model than lot of images.
after training the model
when i ran the last cell
i got this error
NotImplementedError: A UTF-8 locale is required. Got ANSI_X3.4-1968
but it was fixed by running the following code before just before the last cell
import locale
locale.getpreferredencoding = lambda: "UTF-8"
same, did you manage to solve it ?
@@faj6485 solution is mentioned
Hey can u / do u have a video on importing the existing annotations and training for yolo_v8?
Hi, what do you mean with 'importing the existing annotations'?
Hello! I'm having some trouble doing the configuration of the dataset. My code for the yaml file is here below:
path: C:\Ze\\2024\PosIFSC\TCC\Python\Code\dataset
train: train\images
val: valid\images
names:
0: pig
That is the error presented on the prompt.
RuntimeError: Dataset 'C://Ze/2024/PosIFSC/TCC/Python/Code/data.yaml' error object of type 'NoneType' has no len()
I'm using VSCode on Windows 11. I've tried everything but I just can't get it right. Please help me!
hi, I have a question. if I divide the dataset into training and testing only, is it necessary to run the validation part? and if not, during inference how to find out the mAP? plis help me🙏
The content was awesome. But how to run that video in google colab
TY very much for the great video. Edge case question: I'm training my model on only 10 images and I'm never getting any predictions even tho my images are well labeled. Do you have an idea why?
10 images may not be enough, try with more images, a few hundreds at least. 🙌
Very good, thank you very much
You are welcome! 🙌
Isn't the Yolov8 output format of the form 1x8400x84...Shouldn't every image has 8400 predictions of bounding boxes?
Nice video. Thanks)
You are welcome! 😃💪
Extremely useful tutorial for which I thank you. Question please : I looked at your Gti account and could not find the script you are talking about at 1:00:15 ? Thank you. Philippe
do you mean the script I use to create the video with the predicted bboxes on top?
Can I do this for object detection
@Computer vision engineer Will this work for YOLOv5, working on a university projekt where I use a Stereo Vision camera to detect boats and measure the distance and classify them, this video is perfekt for my case. Can I use it in YOLOv5?
Not sure if the training is the same with yolov5, but everything related to the data annotation and preparing the data is the same.
is it possible to convert yolo model to tensorflow2???
where can I find the alpaca videos? I am at the unseen data testing stage.
Take a look here www.pexels.com/search/videos/alpaca/
thanks for the video ! however I followed all your steps and eventually got nothing in the pic with predictions. no rectangle. I'm wondering if it has to do with the number of photos I trained. only used 6... if you have any advice on where the problem could come from, it would be super nice !
Hi, yes it is probably because the number of images you used. Try to collect more images and try again. 🙌
I just want to make sure that I just need the main file and the config file that you have presented, right ? other files (predict.py, results.py...) are not requested, aren't they ? because I also saw other lines of code in your link below so I just don't want to be confused... thanks again@@ComputerVisionEngineer
can i know the source from which u downloaded the image dataset other than the one u mentioned in the video cause i want to train the model to detect Ash trees ,any suggestions would be of great help
thank u
can u help me in training with my custom dataset
Once the model is train, whats should be the hardware to run it?
Hi, if you are referring to an edge device, there are multiple ways to run it. In one of my recent videos I show you how to run it in a Raspberry Pi and a coral accelerator. 🙌
@@ComputerVisionEngineer Oh thx. So some "tops" should be enough in this context. I'm trying to get the correct hardware to train but it's not clear to me to run it on production, maybe I was thinking of the rockchip with 6 tops I'll start from there and see in practice what hardware I need.
Qustion: Does the sharpness of the pictures play a role for the detection object detection?
What do you mean with 'sharpness of the pictures'?
@@ComputerVisionEngineer I mean if high resolution template pictures could make the detection programm working slower or not?
Hi.. I have a dataset consisting of png images(no bounding boxes) and annotations of corresponding images as. xml files(PASCAL VOC).. HOW SHOULD I TRAIN? ALL TUTORIALS I SEE ARE OF JPG AND TXT.. PLEASE HELP
what is the val file in cofing.yaml and how i can create it
Can the dataset annotated be used with yolo v5 instead of yolo v8?
Doing this from a laptop. It uses my integrated GPU instead of my discrete GPU. Any fixes?
Hello sir can you explain why some dataset get 0 accuracy?
It is probably related to the data or the training process.
how to convert .pt format to .onnx format?
Im new, and I always get this problem said 'FileNotFoundError: [WinError 2] The system cannot find the file specified' When i run it on my VSCode. I've made the config.yml file, anyone have any suggestion? please and thank you
okay you have trained the model how do you test it on images it has never come across before?
take a look at the script predict_video.py, also take a look at my other tutorial on object detection + tracking with deep sort, I make inferences with unseen images in that video. 🙌
Hi! I just came across you're tutorial and I love how detailed it is! But if I may ask, do you know how one would do transfer learning in YOLOv8?
Hey, in order to do transfer learning from a model trained in the coco dataset you should create the model like this: model = YOLO('yolov8n.pt') 🙌
Muy buen tutorial amigo
thank u, nice tutorial.
Glad you enjoyed it! 😃
hello sir i have another question, how do i extract the coordinates of the bounding box so i can draw a centroid of the bounding box?
where do i find the .pt file to put the model path?
Buenísimo! muchas gracias!
De nada! 😃🙌
Hi mr. I want to automate palm tree counting using yolo v8. Can the result of the testing export as shp file so i can generate the coordinate from all the detected tree?
I have the file Yolo v8 in format, this file needs to be formatted in my darknet, how can I do this?
i really like your tutorials , but i have an issue during the custom object training i tried it with google colab and also with pycharm but in the pred. image it doesn't show the prediction neither rectangle box nor name tag. please help bro .................
Instead of rectangular bounding box I need instance segmentation, can you make video on instance segmentation with keypoints for custom datasets, please
Hey, please take a look at my other tutorials; I have covered semantic segmentation and keypoints (pose) detection in other videos. 🙌
Why do I keep getting this error. RuntimeError: Dataset 'config_2.yaml' error config_2.yaml 'train:' key missing .
'train' and 'val' are required in all data YAMLs. I use a windows computer. Is there a unique way of specifying the file path in the yaml file?
Hola! una consulta. como puedo hacer para solo descargar la clase alpaca con sus anotaciones y no todas las clases. Saludos!
Hola, te referís para descargar solo las anotaciones de la clase alpaca desde cvat y no las demás clases?
after training where can i find the model file that is .pt file?
runs/detect/train/weights
so cool
😊🙌
Thanks!
You are welcome! 🙌
Hi any idea how this task can be done :When working on annotating lines that connect the light sources , i used polyline tool of cvat , annotated all the pairing correctly , downloaded it in yolo format but the text files are empty, which was not the case while drawing boxes around light sources.Any idea where i might be going wrong or how to save annotaions for polylines to detect the lines in image.
So how should this task can be done?
I searched many ways but couldn't find any way please look into it .
Thank you.
Hey, for polylines you may need to download your annotations in a different format, for example in cvat format. I think it will be very challenging to detect lines using an object detector like yolo. It sounds like a very niche use case, take a look at papers and other resources online, from the top of my head I cannot think in any way to detect lines using an object detector.
@@ComputerVisionEngineer so can I use cvat format to train a model ?
@@ComputerVisionEngineer Thank you for your response , So any idea like how can I detect lines joining light sources in a floor plan as using yolo will not work as I cannot get the annotation coordinates?any other method you think that should work for this task or any model that accepts cvat format?
The video is good and simple EXCEPT FOR the fact that you used the same data for train and validation. Don't do that! That is arguably missleading. It wouldn't be that hard to teach your audiacne what the point of train, test and val is and using the same data for train and val is veerry missleading and confusing since it dosn't really give acurate results.
Thank you so much for your feedback. The only reason I used the same data for training and validation was for simplicity, because I wanted to focus on the overall process of training an object detector. I see your point. I will definitely take it into account for future videos. 🙌
HELLO SIR!!! GREAT TUTORIAL.
I am creating an android application, to detect the bounding boxes and getting the coordinates, how can i do so, can you help me out!!
Hey, if you want to run the object detection natively from your android device take a look at tflite. 🙌
@@ComputerVisionEngineer heyy thanks for replying, but i have converted it to tflite model, and how can i use it in the android studio?
Hello, are there changes you made in the predict.py file regarding the H, W, _ = frame.shape
???
Because I'm getting an error saying
AttributeError: 'NoneType' object has no attribute 'shape'
Your help would be highly appreciated... Thank you.
Hi, that error is probably due to the video path you are trying to predict is incorrect. Make sure the path is correct and try again. 🙌
Thank you very much for that.... I was able to fix it and it worked
I really appreciate
@@leedixon614
I have the same error, could you help me fix it?
Thank you very much for a great tutorial! Great explained and easy to follow. I've now trained a custom model (in Google Colab) for 3 classes. When testing it with unseen pictures, I get good results. If I were to improve my custom model, how do I do that in Google Colab, change the "model = " part and point it to my custom .yaml file? What should I do with my pictures in the train and val folders, keep the old ones that I trained the model on and add the new ones, or empty them and only add new pictures? Should I also add new pictures to the val folder?
I am glad the video was useful! 😃 Keep the old pictures you trained the model on and add new images, yes it is convenient to add images in the val directory too. 80/20 is a good ratio of images in train/val, so if you add images in the train dir also add pictures in val so you maintain the ratio. 🙌
@@ComputerVisionEngineer Thanks! Worked like a charm!
Another question, if you happen to read it 🙂
I'm thinking of using YOLOv8 to look for defects on an object. How should I go forward with labeling and which classes to use?
Should I use one class to put a label around the whole object without defects -> class "OK_object"? What about when it is defect(s) on the object, should I then label only the defect(s) -> class "Defect" and not label the whole object in addition?
Or should I just label defect(s) and not label obejct(s) where no defect(s) is found, so that objects without defect(s) will be 'background' when training my model?
can we apply yolo v8 on low light data?
Sure.
Many thanks!
😊🙌
Is that alpaca_detector.pt in your script last.pt?
Yes.
How do i create the .yaml file in google drive
Hi Felipe , I have run into problem, since I successfully compiled my code but unfortunately it wasn't showing the *runs* folder which leaded to the results, plz help me out.
Hey, have you been able to fix it?
@@ComputerVisionEngineer Yup I fixed it, All that time I saved the Folder *Label* 'L' as starting letter as capital instead of small letter 'l' which I should have done. But I corrected it and now I am gonna start working on the License reader from one of your videos on the same 👍
@@TheConservativeKnight6809 awesome!
can anyone tell me who to test image on this?
Heyaaa! It's an awesome tutorial! I have a doubt though... now I am getting the output MAP50 for all the classes as a whole. But in my project, there are two classes, how to get the individual mAP50 for each of them? Thanks in advance:)
Hey, if it is not provided in yolov8 results you may need to compute it yourself. Take a look at this repo github.com/Cartucho/mAP.git, it seems exactly what you need. 💪💪
where i can get a full source code for Visual Studio?