Тэги:
#padding_in_cnn #types_of_padding_in_cnn #convolutional_neural_network_matrix_multiplication #same_padding #padding_same_keras #padding_in_machine_learning #convolutional_neural_network_equations #zero_padding_in_cnnКомментарии:
thanks Sir
ОтветитьThank you man clear and simple as it should be
Ответить🤩🤩🤩🤩🤩
ОтветитьStrides explanation?
ОтветитьA small question : Are filters always a square matrix or their dimensions could be adjusted by user to their use ?
ОтветитьGreat Effort to make the things so simple. Dedicated teacher. Keep going...
Ответитьsir why should we take always filter is 3x3
Ответитьfor a given input and output , can we find the corresponding filter
ОтветитьAre the values inside the filter always the same , and how are they computed?
ОтветитьGood teaching 🤠
ОтветитьGreat Bro...! But i think Binary value '0' means Black & '1' means White. is it right sir?...
ОтветитьThank you, Krish sir. Nice concept. Beautifully explained.
ОтветитьCan we use 2*2 filter
ОтветитьIn machine learning, there ANN. Also in deep learning, namely in CNN there also a classifier ANN. How can we pass numeric data (.csv file) as an input to a convolutional neural network?
ОтветитьWorthless without code
ОтветитьThanks Krish
ОтветитьYou might got 10 dislike mistakenly bcoz while watching this video I have unintentionally clicked the dislike button but while upvoting it I get to know and rectified my mistake 😅
ОтветитьEasy and to the point 👍
ОтветитьMissing denominator as stride in the formula because your are taking stride one bit what if stride is two
Ответитьthankyou sir..
ОтветитьAwesome video! Very explanatory, thanks.
ОтветитьExcellent
ОтветитьAwesome sir
ОтветитьGreat video. can you please let me know what mike you are using?
ОтветитьYou forgot to mention that the strides may also have an effect. Your example is true if the stride is 1.If the stride was to be 2 for instance, things would change. Use the formula;
Dimensions = [(input image size-Filter size+ 2*Padding size)/stride size + 1].
In your last example with the stride of 1 and padding of 1 as you only have one layer of 0's surrounding your original image,the solution is worked out as follows:
Dimensions = [ (8-3) + 2 * 1)/1 + 1]
= 8
Therefore the feature map has dimensions of 8 * 8
Thank you so much sir for this easiest explanation.
ОтветитьAwesome 💞
ОтветитьThank you Sir
ОтветитьExpecting videos on 3D CNN as well ...
ОтветитьThank you sir
ОтветитьVery well explained! Kudos Krish!
Ответитьin 6×6 matrix if stride =2 and padding=0 what will happen
Ответитьgreat video.Please provide the ipynb notebook link wherever necessary.Thanks.
Ответитьheang natuuuuuu
ОтветитьSir , one ques: when we can use convolution so it's okay to be if we've lower pixel size compare with input , losing data and padding could see it's necessary to make sure not loosing any data.
ОтветитьSir please add RNN lectures also
ОтветитьNext video on Recurrent NN , GANs,Autoencoders , Natural language processing please
ОтветитьPlease upload more videos in this playlist :) eagerly waiting
ОтветитьHi Krish Sir, I really like your teaching style.... you have made clear many many doubts of mine. Thank you very much. And Sir I have a request, kindly upload the further videos on this series as I have got exams recently. God bless you .
ОтветитьWhen are you uploading a new video on deep learning ????
ОтветитьHi Krish....cud u continue sessions on GANs,Variational Autoencoders,CNN and its various architectures and applications in img processing like Unet for semantic segmentation......
ОтветитьPlease give videos of practical implementation also.
Ответитьplease upload more videos on deep learning.....waiting for more videos
Ответитьhey krish do upload lectures on RNN,Autoencoders , boltzmann , belief networks and GANs too with implementation
ОтветитьKrish,
Please do more videos on CNN,RNN .Waiting for it
sir please upload the videos regarding recurrent neural network
ОтветитьHello Sir..! can you upload a practical example that how to build an Image Classification Model in deep learning??
Ответитьkrish kindly do video'S on RNN ...very much appreciated for your effort
Ответить