Autoencoders with keras-core using the torch backend

This notebook illustrates the usage of keras-core with torch with an example of a small yet efficient autoencoder on the MNIST dataset.

It was originally run on a laptop with:

• Intel(R) Core(TM) i7-1165G7 @ 2.80GHz
• 16GB of RAM
• No discrete GPU

Scripts and Notebook

You can find the notebook here.

You can find also find a training script and a testing script under ./scripts/.

Dataset

I used the MNIST dataset in npz format from here: https://www.kaggle.com/datasets/vikramtiwari/mnist-numpy?resource=download.