How to get the function 'convolution1dLayer' ?

I haev been working on the 1D CNN. But in matlab if i try to use the function 'convolution1dLayer'. Im getting a response as,
>> layer = convolution1dLayer(11,96)
Unrecognized function or variable 'convolution1dLayer'.
Did you mean:
>> layer = convolution2dLayer(11,96)

5 Comments

I have tried to update. but I'm already using the laterst version, R2021a.
R2021a is not the latest version. Release R2021b came out in September.
Yes, I have installed the R2021b by uninstalling the previous version. Is there an any other way to change the MATLAB verison without unintalling the previous version ?
You can install in another path and use both the versions.

Sign in to comment.

Answers (2)

There are any solution?

7 Comments

If you're using a release prior to the introduction of that layer in release R2021b, the solution is to upgrade.
If you're using release R2021b or later, check the output of ver to determine if you have Deep Learning Toolbox installed. If not, install it (purchasing it if you don't already have a license for it.)
Steven Lord.
I have the R2021b version upgraded and the license is in order. The "Deep Learning Toolbox" is installed and even so the command "convolution1DLayer" not appears.
Thanks
What does this command show when you run it in your installation?
which -all convolution1dLayer
/MATLAB/toolbox/nnet/cnn/convolution1dLayer.m
As it can see, convolution2dLayer is there, but convolution1dLayer not.
C:\Program Files\MATLAB\R2021a\toolbox\nnet\cnn
>> which -all convolution1dLayer
'convolution1dLayer' not found.
>> which -all convolution2dLayer
C:\Program Files\MATLAB\R2021a\toolbox\nnet\cnn\convolution2dLayer.m
Maybe i can download it or installed manually?
You said "I have the R2021b version upgraded" but the path you showed indicate you're using release R2021a: "C:\Program Files\MATLAB\R2021a". You must use release R2021b or later to be able to use convolution1dLayer.
Oh, Maybe I was wrong thinking that I updated. I will confirm this.
Thank you very much!

Sign in to comment.

yes,sir,may be use 2D to replace 1D,such as
imageInputLayer([1024 1 1])
ans =
ImageInputLayer with properties: Name: '' InputSize: [1024 1 1] Hyperparameters DataAugmentation: 'none' Normalization: 'zerocenter' NormalizationDimension: 'auto' Mean: []
convolution2dLayer([100 1],3,'Stride',1)
ans =
Convolution2DLayer with properties: Name: '' Hyperparameters FilterSize: [100 1] NumChannels: 'auto' NumFilters: 3 Stride: [1 1] DilationFactor: [1 1] PaddingMode: 'manual' PaddingSize: [0 0 0 0] PaddingValue: 0 Learnable Parameters Weights: [] Bias: [] Show all properties

2 Comments

Thanks yanqui liu, the problem is my data.
I have 1D vectors and i want to make a CNN with a BiLSTM, so i must use de convolution1DLayer.
yes,sir,may be use sequenceInputLayer to get model,such as
layers = [ ...
sequenceInputLayer(numFeatures)
lstmLayer(100,'OutputMode','sequence')
dropoutLayer(0.3)
lstmLayer(50,'OutputMode','sequence')
dropoutLayer(0.2)
fullyConnectedLayer(numClasses)
softmaxLayer
classificationLayer];

Sign in to comment.

Categories

Products

Release

R2021a

Commented:

on 31 Jan 2022

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!