3:00 the rotation example you use would actually confuse a CNN just as much as a transformer. Convolutions are translation invariant, they are *not* rotation or scale invariant without modification. Just want to point that out. Otherwise though excellent video.
Thanks for your feedback Shravan! I mostly use Google Slides. I have created my own template for AI BItes. Every now and them I use Manim which is an amazing math animation library written in Python. Hope it helps.
Thank you for that video. I've been studying this model and helped me a lot.
Glad it helped 👍
super interesting, first time I saw a video by you, directly subscribed! Thank you
Thank you Wil. Very encouraging to continue my work 😊
Thank you so much for the brilliant explanation !!!
I really wish you had explained the self attention equation too.
Will try and make one about it. Thanks 👍
3:00 the rotation example you use would actually confuse a CNN just as much as a transformer. Convolutions are translation invariant, they are *not* rotation or scale invariant without modification. Just want to point that out.
Otherwise though excellent video.
yup I should have mentioned a better example. I remind myself the whole idea of RotNet now :)
Thanks for all your work, your explanation is impressive with such beautiful illustrations.
What tool do you use to create those illustrations?
Thanks for your feedback Shravan! I mostly use Google Slides. I have created my own template for AI BItes. Every now and them I use Manim which is an amazing math animation library written in Python. Hope it helps.
Great explanation, please do (Slimmable Net) next.
yup sure will try and do Slimmable Nets next. Thanks.
Nice informational Video, keep going
Thanks, will do!
Your videos are great, congratulations and thanks a lot for doing it!!
Thanks for your compliments.
Can you take about CoCa(Contrasive Captioners) and how to use it
Sure will do at some point. Thanks for watching the other videos meanwhile :)
thank you so much it is great explain and i am looking for that code to implant it in my dataset plz if u do it
will try and do implementation videos at some point in time. Thanks for the constructive feedback
I would say Conv with self-attention🙃
Similar to CNN+BERT
Dude, do something with your mic. It is on and off during your video. Sometimes it gets off when you say something very important.
Hey thanks for pointing out. Let me check the setup again and see what's the reason for the intermittent disturbance.