How to Design a Neural Network | 2020 Edition

[@saisumanth94]

There was a moment where I felt I was lost & concepts became overwhelming, your videos guided me through all the obscurities and helped me understanding concepts clearly. Great explanation, great presentations. Keep it going bro

[@jennajulit6895]

This video is so great that I realize that I don't need to start from scratch :)

[@Eric-hb9rm]

6:48 This is a U-Net architecture right? What is special about it? Thank you

[@theeFaris]

Why doesn't this have more views? Fantastic

[@__init__k917]

Dude!!! you are superb, can you please point out good resources from where to learn all these and the underlying maths.

[@kdsuch]

Great work and happy new year! :) Would like to see a video about deep learning with small data sets :)

[@faisalalsrheed82]

Happy New year. Thank you for other great video.

[@airepublic9864]

Make more, you satisfy my curisity

[@X_platform]

Happy belated New Year! :)

[@itayblum3405]

Well done man... what a great video

[@artinbogdanov7229]

Another great video. Thanks for doing it!

[@amirpace]

Hi, I want to Design a light weight CNN for Masked Face Recognition as the traditional Face Recognition Algorithms have a performance degradation of up to 70% just like mobileNet. Can you suggest me how should i go about designing it?

[@hakanaktas2665]

Great work, CNN architectures has already high accuracies. There are many studies sppeding up CNN architectures and offering Hardware friendly architectures. I think these studies will continue untill deep learning is efficiently applicable in many embedded devices.

[@leoisikdogan]

Thanks. Indeed, efficiency and hardware friendliness of architectures gain importance as on-device deep learning applications get more prevalent. Recently, my colleagues and I also published a paper in that domain: kzbin.info/www/bejne/nou6gmWIqKtqfNE

[@billy818]

This is really great thanks. Also happy newyears :)

[@haohuynhnhat3881]

Love your videos very much

[@ll2pakll]

He says I can SELECT an already working neural network. But where do I choose them? Where do they lie so they can be selected?!

[@leoisikdogan]

You can start with the state of the art methods here: paperswithcode.com/

[@NBAyyy]

Great Content. Love this~

[@user-wr4yl7tx3w]

can you do a video on transformers?

[@ThamizhanDaa1]

great video!!

[@mattipiirainen7440]

I stumbled upon this video and all I hear is familiar words in unrecognizable patterns. :D

[@mache-xw7ox]

Great video!!! Is there any paper related to this topic??

[@leoisikdogan]

Thanks! Yes, I put the links to the referenced papers in the description.

[@abd_sam]

How can I be so late? Thank you for making this video.

[@sureshothwarang]

Do you have udemy classes? I'm buying! I want some object detection classes please

[@leoisikdogan]

I don't have udemy classes but I might make a free video on object detection. Thanks for the idea!

[@pervezbhan1708]

@@leoisikdogan please donot be like Siraj, we are learning from you free

[@XX-vu5jo]

@@leoisikdogan hoping for that soon! OpenCV and SSD, YOLO, and other ways to implement and train would be very helpful. If its paid I'll still go for it. You're really good doctor! Will try to cite some of your works in my future papers. ;)