Object Detection using YOLO - Getting Started

[@oliesting4921]

Great! so glad I have found this channel. Your work and YT lectures are much needed in the field of DS/ML/AI. Thank you.

[@AIEngineeringLife]

You can check my repo here and folder within it - github.com/srivatsan88/KZbinLI

[@Augmented_AI]

Great explaination 😊 well done 👍

[@AmanKumarSharma-de7ft]

This is a life saver sir. Thanks for the upload and I will be waiting for the custom training video using yolo v4.

[@Cricketpracticevideoarchive]

Sir I can’t tell you how glad am I for this video, if you are marking a series of this then it would be extremely helpful

[@AIEngineeringLife]

Yes..I will be doing couple more videos on Yolo. Yolo installation is not straight forward so in this video wanted to introduce installation and pre-trained models

[@rushirajparmar9602]

Thankyou sir for this!! I hope you will be covering variants of YOLO in this series as YOLO V4 is already released. Frankly speaking this channel covers all the practical implementation in a nice lucid manner! Love this approach! Especially your videos on AutoML. I would love to see more videos on End to End ML model Deployement on AWS or GCP! ❤️

[@AIEngineeringLife]

Rushiraj.. I am covering Yolo v4 only in this. Can you please elaborate on variant as I will be covering few more from Yolo.. On model deployment in GCP you can check my below playlist - kzbin.info/aero/PL3N9eeOlCrP4VXtFJTjmGsqI-Emk2keVL AWS is in plans in future

[@rushirajparmar9602]

AIEngineering nothing specific sir what you are covering is great I just mean we have Yolo tiny even some SSDs and other architecture so which option will be best in comparison to YOLOv4 if I want to deploy this model on edge? Or if YOLO v4 is that fast enough I think the issue is solved!! Btw looking forward to your AWS series!! Thankyou!

[@bunnyvlogs7647]

Great sir. .ur awesome... please do series of videos on this

[@AIEngineeringLife]

Thank you and Sure Sai.. Will continue with Yolo post my Time Series video series ..

[@abhishekpawar921]

Amazing tutorial!

[@AmanKumarSharma-de7ft]

Sir I have a question though ....if I am using windows, then for what os is the exact cudnn 10.1 do i have to download? the Linux or the window one. and if Linux hen which architecture ?which distribution ?which version? there are so many cuda drivers that i cant decide what to go for. please suggest. I don't have a GPU in the local and is currently running on windows

[@AIEngineeringLife]

If u r having windows and no gpu then I can ignore the makfile changes step and directly call Make command. You can check darknet repo for install steps

[@AmanKumarSharma-de7ft]

@@AIEngineeringLife sir as i will be working in colab which doesnt have any possible connection with my local system so cant i download the cudnn linux version u were using and upload it to gdrive? Will that not work?

[@lilasrinivasreddy]

Can You make a video on how to convert these weights into tflite quantized model and deploying into mobile applications?

[@AIEngineeringLife]

Sure will do.. But in case if you have not seen I have seperate videos on TFLite but for image classification problem

[@valerysalov8208]

Can Tranformers or BERT variants be used for metadata extraction?

[@AIEngineeringLife]

For object detection?.. I did not get..

[@valerysalov8208]

@@AIEngineeringLife Text Extraction

[@roshanshaji]

Plz include custom training using our own image set!! Its a request 🙏

[@AIEngineeringLife]

Roshan.. Yes it is in the plan..

[@vishnuk.v8529]

@@AIEngineeringLife when will it be posted?

[@sdhilip]

Hi, Good video. Thanks. Could you please share the notebook?

[@AIEngineeringLife]

Here you go - github.com/srivatsan88/KZbinLI/blob/master/Yolo_Installation_and_Prediction.ipynb

[@sdhilip]

@@AIEngineeringLife Thank you

[@helloai3470]

Sir, first of all thank you for this amazing video, very informative!!!.... I have 1 question,, I am using colab for building emotions detection model using mobile net and open cv,,, however to train,, I need to upload kaggle dataset of images to my Google drive,, Data is 60mb cropped,,,, but it is taking so much time to upload to the drive,, m stuck,,, is there any other faster way to load dataset into colab???? I really need your help sir... Thank you, Rajeev

[@ramanjaneyuluthanniru1428]

use kggle API automatically it will load data set in your colab drive

[@AIEngineeringLife]

You can use Kaggle api as Ram has mentioned and directly pull the dataset from kaggle to colab and from there copy to drive

[@helloai3470]

@@ramanjaneyuluthanniru1428 thnx ,,will do that...

[@helloai3470]

@@AIEngineeringLife Thank you sir, I am trying now...

[@sainikhilesh3467]

Sir great tutorial..it was very informative. Recently I tried this but I could make detections on video after loading them in colab and then predicting, however can you please let me know how to implement it in REAL TIME? Any modification that can be made to do this in the .ipynb file in your GitHub repo. Thank you so much in advance.

[@AIEngineeringLife]

In colab it is going to be difficult for real time. We can access webcam in colab but not sure if we can get frames continuously... I will try it out and see

[@valerysalov8208]

How to make your dataset gold standard?

[@AIEngineeringLife]

Frankly I dont know what gold standard means.. I would say if dataset is good enough to correlate events from multiple sources and generate good insight then it will be good learning source

[@ramanjaneyuluthanniru1428]

Sir, Please upload the code in github...for practice purpose we need it.

[@AIEngineeringLife]

Here you go - github.com/srivatsan88/KZbinLI/blob/master/Yolo_Installation_and_Prediction.ipynb

[@ramanjaneyuluthanniru1428]

@@AIEngineeringLife TQ