Question for you: The first time we trained the model we got a 94% accuracy (16:52) but the second time we trained the model we got a 96% accuracy (18:36). Both times we are using exactly the same dataset and exactly the same classifier, why are we having different results every time we train the model? 🕵‍♂ The answer is available to all the Computer Vision Enthusiasts and Computer Vision Experts of our community here: www.patreon.com/posts/image-feature-88454744 🙌

Thank you for this video, I am currently working on my thesis and content like yours helps me understand topics more quickly. God bless you.

almost 10k, congratulations felipe

@ComputerVisionEngineer

10 ай бұрын

Thank you!! 😃🥳🥳

Nice...

Thanks for this tutorial. I got score 1 with the same dataset!!!

@ComputerVisionEngineer

9 ай бұрын

Awesome!! 😃🙌

Hello! Can you please help me: where should i store images for the classifier (if i share model with someone)?

Usedful ❤❤❤

@ComputerVisionEngineer

10 ай бұрын

Glad you enjoyed it! 😊🙌

I love the added dramatics 🤣

@ComputerVisionEngineer

7 ай бұрын

😂😂😅

Thank you

@ComputerVisionEngineer

10 ай бұрын

You are welcome! 🙌

Hello, Thanks for the video. Your contents are super. As a beginner, to computer vision please guide which video from start and which one mid and which one end from your videos. If possible can you make a playlist on your channel thoose who need to start from the beginning? Thanks again...!

@ComputerVisionEngineer

10 ай бұрын

Sure. I have made some playlists already, but I will create another one with a sequence of videos ranging from most beginner to most advanced. 🙌

This is a channel where every comments gets a reply

@ComputerVisionEngineer

10 ай бұрын

I try to answer every comment, not always succeed though 😅

Where can i find the train file, and val file please ?

May I ask what is your OS you frequently use for computer vision work? I recently tried an image classification and need to config tensorflow for GPU because training on my CPU (~2.40 minutes) takes a longer time compared with training using GPU (~1.30 minutes). The update from tensorflow says that it requires wsl for GPU setup on Windows and MacOS is not supported yet.

@ComputerVisionEngineer

9 ай бұрын

I use Linux Mint in my local computer. When I train a model in the cloud, the OS is usually Ubuntu.

@nguyenthehoang9148

9 ай бұрын

Thanks @@ComputerVisionEngineer Have you considered a video to share your experience and opinion regarding training models on the cloud? The training has some cost, but if the cost is reasonable and gets the work done (even for learners), it's worth it.

@ComputerVisionEngineer

9 ай бұрын

@@nguyenthehoang9148 yes, good idea, I could make a video about the different ways to train a model. Local vs Google cloud vs ec2, for example.

@nguyenthehoang9148

9 ай бұрын

Great@@ComputerVisionEngineer I just checked the pricing of AWS, and it doesn't look too bad. More importantly, I think people who want to pursue computer vision will eventually come to the point of using the cloud for model training.

@nguyenthehoang9148

9 ай бұрын

@@ComputerVisionEngineer I'm doing an exercise with the bird classification dataset on Kaggle, and there are several configurations I want to try. However, it's not very efficient to do it on my local machine; therefore, a cloud might be a must-try solution.

i am beginner... maybe you create tutorial on yolo8 image class very detail, from anotation untill model running...some video i saw only train and val model ...

@ComputerVisionEngineer

7 ай бұрын

Hi, I made a video on image classification with yolov8. Make sure to check it out! 🙌

Can you similar tutorial with Tensorflow ?

@ComputerVisionEngineer

10 ай бұрын

I will try to. 🙌

Hey can you create a tutorial for smart Surveillance which detect human intrusion classify between authorised and unauthorised intruders and notify security person usign YOLO V8 please

@ComputerVisionEngineer

9 ай бұрын

Hey, sounds like a very cool project! I will try to do it. 🙌

@lawrencerodriques7153

9 ай бұрын

@@ComputerVisionEngineer can you do it in this week please 🙏🙏

Why is it rare to find a object classification built with pretrained models like yolov8 and yolonas?

@ComputerVisionEngineer

10 ай бұрын

The reason may be that yolov8 and Yolo nas are novel technologies. The library I used in the video, Scikit learn, it is been around for years. 🙌