Quick Deploy: Object Detection via NGC on Vertex AI Workbench Google Cloud

[@fatihtarikata4607]

Very underrated video, Ahmad is pushing Google products, I mean this guys should be recommended everywhere.

[@ahkajans9017]

I am in MLOPs and Google’s Vertex AI is one platform-as-a-service for machine learning and AI developers to encourage MLOps in the industry that is super highly recommended.

[@ahmedalahly.403]

I used to use openCV to extract segments from vdo, but this video is on another level.

[@minecrftlegend5116]

Highly recommended video! A hidden gem! Suggest beginners to pay attention it will save time.👍👍👍

[@محمد_العراقي-ب1ر]

you and your videos deserve more attention than that. you are super hero of engineering student. thank you:)))))

[@ibrahimglr.16]

Sir you are awesome. i regret not visiting this channel earlier. Thanks a lot sir.

[@twitchssbaba1637]

Just getting started with Google Cloud. This video was a great help. Thanks so much.

[@cagdasulucanl6371]

Thanks again for taking the time to post such detailed videos..

[@gizemshn5974]

What an underrated professor !

[@ramin29]

Hi there, finally the best tutorial combining Python's OOP and Tensorflow w/ Object detection. Big Thank for that! Very happy to see VS code instead of Jupyter ...

[@username56561]

Thank you for sharing your work with us Ahmad !

[@fghnjgthnyhjh3240]

Sr you are the best!!!! your video is just like i need. In a close future i will signed in your courses

[@clgnoyunyt7306]

tremendous upload Ahmad. I broke that thumbs up on your video. Keep up the brilliant work.

[@moviesok3326]

Very good lesson. Very well-produced.

[@iiwiqhwjjaia3400]

Thanks for the tutorial is very clary! From Patagonia 👏

[@joikaygamers4405]

Thank you very much for this wonderful tutorial it was great help .👏👏😊❤

[@zeonedits814]

Very well explained. Thank you!

[@cnaryuksel2208]

Your great content needs to be showed for bigger audience! This of course takes time especially given the biased youtube algorithm.

[@wraqrwqrwq1222]

Excellent tutorial. Very neat and straight up to the point.

[@Saliicii]

Advice: Google cloud vertex AI it automates things for you and lets you focus on more important things in a machine learning workflow. With Vertex AI, Google targets the newbies to provide ease of use and the experts to save time for basic tasks.

[@yonxfb3195]

Yes. Custom object detection with any of the 40 models tutorial is planned for next week.

[@emrebeys3708]

Start watching you and you are awesome ✨ ❤

[@uejsjedjejjsjs2954]

Ahmad Bazzi is a God !

[@PhucHoang-xe1hp]

nội dúng sáng tạo quá hay rất bổ ích rễ hiểu

[@huuhh3113]

Thank you for the appreciation.

[@gorkemkas6668]

These lectures are really amazing so far. Thank you so much, Sir.

[@lilhase878]

Great video!! Why don't you use pyinstaller for exe packaging?

[@_echomoony_1290]

Thank you so much sir learning so much from you :)

[@editizmft]

Vertex AI changed my life ..

[@wbcarftworld8582]

Now this is what i call, A Unified Machine Learning Workflow.

[@gokdenizakcalar6043]

Excellent Ahmad, thank you !

[@muzikdunyasi314]

Your video is really helpful! How is the AI platform different from object detection containers and then set firewall to access it?

[@memotibirbesdokuz1447]

Ahmad Bazzi - I am looking forward to this :) Another question: Could you upload the code you've written in the video (e.g. to github)?

[@mehemmedcebrayilov112]

Nice vid on object detection.

[@emirgul2527]

GREAT WORK! I waited for python to .exe Thanks:)

[@benyigit9030]

For those who dont know, Vertex AI is a managed machine learning platform that provides you with all of Google’s cloud services in one place to deploy and maintain AI models.

[@klklklkklklkl7868]

deliated convolution tricks used in deeplab models can detect certain features at different scales.

[@talhady1687]

Amazing video thanks a lot!! : )

[@pratiksintv226]

I am glad it was helpful.

[@DD-gy5dg]

this very helpful for me :v thank you so much

[@cahitgamer]

Will probably do a custom video on object counting soon!

[@coolpix8890]

Why doesn't KZbin show this to more people ?

[@umutcuk31]

great video, kindly make a video how to deploy object detection model trained using detectron2

[@caglarmetin1144]

Great tutorial!!

[@emiragar4799]

Excellent video

[@siracdinc1723]

Thanks for the awesome video, and im trying to run it on a Tegra. Any idea how to do this ?

[@pragmatic6691]

good course, thanx

[@alipolatceylan5213]

very underrated.

[@jkfghjdftg4739]

Good tutorial. I'm getting duplicate very similar bounding boxes on all of the objects I detect and I'm not really sure why and how to fix it.

[@topraktogg7803]

Great explainer

[@marstrapbeat7693]

Oh God - how does Ahmad manage between research and video creation with god level content ?

[@chosonne5925]

This one was great. is it possible for a training tutorial for custom object dataset by using models from Model Zoo? like faster ccn? building up on this tutorial would be great

can we adopt same procedure for deploying object detection model trained using detectron2

[@mizahsorkral1533]

very helpful video

[@papatestapates3411]

If the covnet is trained for a specific size of the image, how will it predict 0 or 1 for a different size of the image when we increase the size of the input image ?

[@kralbilgiler9911]

Vertex AI is a managed machine learning platform that provides you with all of Google’s cloud services in one place to deploy and maintain AI models.

[@efedasdemir5687]

Hey, what is that resnet50... file you have used??

[@user-lg2nh5rl1i]

Do the course have any perquisites apart from python?

[@the.montajer1338]

I'm running TF on CPU, properly installed, works as fine w/ other codes but I encountered the issue at 17:18, haven't see that in any comment in here so hopefully u can help. First iteration of loop is OK and returns coordinates of bboxs (i kept print) but cannot put values into list, I got an issue: 'float' object has no attribute 'tolist' ... numpy is working properly, tried separately, tried bboxs as other numpy float types ... no way. Any help appreciated. I can definitely make different for loop to get the results but this one from your code seems to be very convenient. THX!

[@npc2600]

This is awesome, thank you for sharing with us! What if we would like to add our own images like in transfer learning? For example, drones are not in that list, could we detect between 3 different types of drones (Phantom, Parrot, Tello) and disregard the other classes from that list?

[@ryanclips7068]

I've tried different python to exe converter in the past and cx_freeze is the one that always performed the best and it's able to convert complex and full software into exe.

[@selimcangunduz772]

Hi Ahmad. I want to draw a rectangle around the crowded place in real time and I want to measure the square meter of the area I draw. But I couldn't do it. I would like to get your opinion on how to do it. Thanks..

[@tahaylmaz91]

This guy only has 146K subscribers ? Wtf KZbin.

[@soccyshine]

How would I create a recognition model for different object classes?

[@nisasblog4163]

I also had the problem and solved it this way.

[@alizeynalov6758]

Can we use ResNet with BigQuery, i am working on a project where i need the latter.

[@emgarage2185]

HEllo please how can we do our own implementation of Faster CNN without using pre-trained models?

[@jeferson1257]

I had a look at NGC a few months ago, but had got quite confused. Your video is really clear and helpful! What is the advantage of using notebook instances in NGC when Google Colab is available for free (I do understand that Colab is slightly different from the JupyterLab interface when working with notebooks)?

[@kullanmiyorrr8403]

underrated.

[@ares9170]

Hi Brother, I need help to select between Tensorflow & OpenCV, what should I do. And I need some consultancy for my CCTV project.

[@mehmetdag4530]

can i try this on a raspberry pi or jetson nano?... currently working on a personal AI companion.

[@berkyagc4035]

Glad to hear that ;)

[@xvortex8482]

I am using CNN's with a lot of layers. They use padding so that the input size doesn't shrink. This makes this approach not so straight forward. Any idea how to deal with that? Another case are Resnet like blocks, which have different convolutions on different paths merging. Without padding this is difficult, any idea?

[@hugovegeta4012]

thanks for video. So, I want to do objet tracking and coordinate in OpenCV. What advice me, How you can help me tot do. thanks.

[@jeffreybajrami8899]

I really like your videos :) Are you planning to do a video on tensorflow keypoint detection (with custom data) or just custom data with tf model zoo (like with detectron2)?

[@150a-z8d]

Is VS C++ must be installed?

[@albumuzik9145]

Hi, with this line: cv2.putText(frame, str(class_id) .... ) I get the class printet on de box but in square brackets [0], so I cant print the class name. What I am doing wrong?

[@trikdantips7466]

very helpful

[@aky-alikemal7754]

How did you do it ? Can you guide me ? I can not use my credit card for the 300$ credits on google cloud.

[@ereneren2788]

My concern is how to protect my custom dataset from being copied or used without a license?

[@ardaozkan3392]

for some reason even after running the pip install opencv command, the dnn module doesn't seem to have the readnet function, is there another way I can get this? please help

[@alperenekici4588]

Do you know how to load dataset from kaggle to jupyterlab without download the data to the computer? Because my dataset is too big

[@zulaabisininfannndedesinin9787]

Increase thresholds in NMS function

[@mamiheizm5758]

Hallo mister, i need help to detect fruit with KNN methode for my home work. Do you have reference to my case?

[@xion5834]

Would be cool to identify pine cones using autonomous vehicle, then pick them up and place them in basket on autonomous vehicle. Save me time going around yard picking up pine cones before mowing.

[@ademafk5413]

Why the main file does not start? I copied the dnn_model folder, but it doesn't start

[@cabbar5433]

My question is: how can I use my own trained model with my own labels, but using tensorflowv2 zoo models?

[@ferhatberberoglu7360]

What always has been difficult for me is multiple camera streams and sending those multiple frames to a detection model (such as Yolo).

[@Merto_GT]

Sure can, just ensure you label the new images as well.

[@ratnabhanu]

Excellent

[@heavenboardpyronic1068]

I was trying to implementing YOLOv3 for object detection using python in pycharm. my code is working fine but it's not detecting bounding boxes with it's label which means that bounding boxes code is working, but only 1 objects are detect. at a time all objects are not detected in the frame...I am unable to find the error behind it. I have used yolov3 pretrained models in my code. Can any one tell me what are the possible reasons?

[@icerikfikirciisteyardimede4841]

Pls reply the model was detecting my face as class id 59 i dont know why?

[@totyfutbol175]

Spettacolare!!!!

[@orgudemoda4964]

Is Ahmad the guy in the Stanford University？

[@MaizerAI]

how did u solve it ??

[@furkankaplan772]

For some reason, it does not use my GPU when I am running it on a video.

[@legendonlinereis6400]

Awesome thanks..

[@gokdenizgenc4696]

Custom object detection is planned for next week.

[@aykutelmas7131]

Do you have a way to do it?

[@ilkaydemiray1690]

This one might work with opencv as well, but I haven't tried it.