In case it's not clear, the formula at 1:56 specifically only applies in our case (3x3 normal convolutions with no pooling, stride 1, etc.). The formula at 1:17 only takes into account kernel size, stride length, and the receptive field of the previous layer. Perceive. Relieve. Achieve. Support me! www.patreon.com/intuitiveml
@Xamulporter
2 жыл бұрын
This is great! Thanks for the easy-to-understand explanation!!!
@IntuitiveML
2 жыл бұрын
You're very welcome!
@mbearwman
3 жыл бұрын
I like your teaching style!
@IntuitiveML
3 жыл бұрын
Thank you! 😃
@jenniferjordache4255
Жыл бұрын
Super useful video, thank you! Exactly the intuition I was looking for :)
@geeky_explorer9105
3 жыл бұрын
I work a lot on CNN and now I am trying to understand how CNN works from gradcam perspectives, and u nailed it to understand how receptive field converges from one layer to another
@IntuitiveML
3 жыл бұрын
Thanks! Hopefully you learned something from the video!
@ocamlmail
Жыл бұрын
Thank you so much, this is really fantastic! Btw, can you explain since 2:40 why colors (orange, purple) don't intersect (or mixed) in case we increasing number of layers? PS: I think one should understand it as 3 separate drawings -- first fully orange, second -- fully purple and so on. Is it correct? Because everything on same pciture\layer really confuses... PS2: looked once again -- you name each layer by it's color, but maybe more better would be to draw 3 separate colorful pictures (grids)? Thank you, really helpful!
@alexz1657
3 жыл бұрын
Your teaching style is very good! But perhaps, the fastly writing move harms my concentration. Thank you very much.
@IntuitiveML
3 жыл бұрын
Thank you! What do you mean the 'fastly writing move' hurts your concentration? Are you saying the writing is too fast?
@biomedicalengineering1326
3 жыл бұрын
@@IntuitiveML yes
@darshanjani7595
2 жыл бұрын
Perfect!
@kaviarasu89
3 жыл бұрын
Your content is good but you are too fast so not able to listen properly. Try to explain slowly
Пікірлер: 19