Комментарии:
What exactly are those "features" in convolutional layers ? I get the point of down/upsampling the data using convolutions, but I cant see how those features fit in the network as whole... How does the network determine the loss gradients of different features with backpropagation, and how do you reduce the number features at the end ?
ОтветитьReduce the speed !
ОтветитьCongratulations on the great content. In this example, noisy images are obtained by adding noise to the existing images first, and then the encoder is trained to remove the added noise from the noisy image. In fact, in this example, the encoder is trained to simply remove random noise from any image. In real applications, however, existing images may contain many different types of noise. So how effective could encoders have been at removing noise? In other words, if we want to eliminate noise from the dataset we want to use, how can we eliminate other noise besides random noise using an encoder?
Ответитьthank you from japan
ОтветитьReally nice vids! Thank you
ОтветитьVery useful, tx!
ОтветитьKeep up the good work!
Ответить