Support for Autoencoder made up of Dense Layers

[expeon07]

Hi, is an autoencoder layer with Dense layers supported by emlearn? I am thinking it's the same as an MLP implementation but I am getting errors when trying to convert the model.

Also, is it possible to deploy it on an Arduino Uno?

Thanks!

[jonnor]

Can you provide the Python code for the model? And the error you get on conversion?

In general there are better tools for, such as NNoM or TFLite-micro for neural network models. But emlearn supports some simple cases, and can maybe be expanded to cover a couple more.

[jonnor]

Now that we have anomaly detection examples, it would be nice to add some simple dense AutoEncoders also there.