My question is as follows
model.add(Conv1D(32, 55, input_shape=(10100,1))
model.add(MaxPooling1D(5))
Input is 1 D signal having length 10100,
this input is applied for convolution with 32 different filters, so 32 Different output is generated
(input image [10100,1]) convolution ( 32 Filters, size [1,55] )= 32 Different Output having size (10046,1)
model.add(Conv1D(32, 55, activation='relu'))
model.add(MaxPooling1D(5))
these 32 Different Output is again fed to next convolution layer which has 32 different filters,
so according to me, conv1d will take one input channel (10046,1) at a time out of 32 channels ( which is coming from the output of the previous convolution layer) and generate 32 different output,
so likewise conv1d will take all 32 input channels one by one and generate 32 different output channel for each input channels, so in total 32x32=1024 channels output should be generated,
but Conv1d output is only 32 channel output is generated instead of 1024 channels.
so what is internal mathematics of conv1d function of Keras which generates 32 channels only with 32 input channels with 32 filters size [1,55]?
model.add(Conv1D(32,(3,3), input shape(80,1,1))
Model.add(Maxpooling(3,3))
The input shape 80 is height of input image and 32 filters are used with 3 kernel size then the output features can be calculated as:
Height-kernel size +1 = 80-3+1=78 features
Then next is max pooling layer, it always decrease the number of output features and output size is given by
(Input size-kernel size +2×padding)/stride+1
=78-3+2×0/1+2 =25 features
Hope you understood
My question is as follows
model.add(Conv1D(32, 55, input_shape=(10100,1))
model.add(MaxPooling1D(5))
Input is 1 D signal having length 10100,
this input is applied for convolution with 32 different filters, so 32 Different output is generated
(input image [10100,1]) convolution ( 32 Filters, size [1,55] )= 32 Different Output having size (10046,1)
model.add(Conv1D(32, 55, activation='relu'))
model.add(MaxPooling1D(5))
these 32 Different Output is again fed to next convolution layer which has 32 different filters,
so according to me, conv1d will take one input channel (10046,1) at a time out of 32 channels ( which is coming from the output of the previous convolution layer) and generate 32 different output,
so likewise conv1d will take all 32 input channels one by one and generate 32 different output channel for each input channels, so in total 32x32=1024 channels output should be generated,
but Conv1d output is only 32 channel output is generated instead of 1024 channels.
so what is internal mathematics of conv1d function of Keras which generates 32 channels only with 32 input channels with 32 filters size [1,55]?
- Badges
- Science topic