Latent space auto-encoder

[ziedhosni]

Hello,

I would like to use an autoencoder to move back and forth in the latent space.

It works basically like PCA but in much more efficient way and it is able to get back to the original dimensions.

This phenomenon is exlained in the link below and I found a couple a script written in python.

Is there any attempt to implement a autoencoder in Labiew that I can use?

https://towardsdatascience.com/generating-images-with-autoencoders-77fd3a8dd368

https://github.com/snatch59/keras-autoencoders

Thanks,

[jiangliang]

Maybe you can lookup here:

https://forums.ni.com/t5/Example-Code/LabVIEW-Deep-Learning-with-Python-Keras-API/ta-p/3998133?profi...

You'll need to change the output of python node to a 2D-array or 1D-array and reshape it I soppose.

Also, I'm learning autoencoder too, I did get the modle running on Spyder, but not yet transfered these things to LabVIEW yet, but I certainlly will because I can use LabVIEW to get the data for this modle.