TFT model output latent space embedding for classification

[Mpaperlee]

I'm new in darts, and I was train a TFT model using my clinical dataset. I was wondering, is their some way to extract the latent spaces embedding of samples? Than I can use these embedding to do clustering.

[madtoinou]

Hi @Mpaperlee,

This feature request is quite similar to the one mentioned in #1649, it's not available yet in Darts.

If you would like to implement this feature or a workaround for your use-case, you can find an example of how to extract some weights from a model here.