1 min readApr 25, 2020
Alexander Graham thank you for reaching out.
By definition, an Autoencoder is built to reconstruct the input. That is, the model has to predict X given X as input. Therefore, model.fit() has X for both input and the output.
Alexander Graham thank you for reaching out.
By definition, an Autoencoder is built to reconstruct the input. That is, the model has to predict X given X as input. Therefore, model.fit() has X for both input and the output.
Director of Science at ProcessMiner | Book Author | www.understandingdeeplearning.com