Sample_Inference for pre-trained VAE or KNN model

[fmarinb]

Thank you for this fantastic tool. I have been able to successfully perform several analyses ranging from unsupervised VAE analysis to creating supervised models and obtaining Residue Sensitivity Logos.
I am currently doing a validation phase with new data, and I was interested to know why there is a way to implement the Sample_Inference() function on a pre-trained VAE model, in the same way that you can with Sequence_Inference(). On the other hand, if a KNN model for repertoire classification is generated from the features learned from the VAE, is there a way to apply that same model to a new set of repertoires?
Kind regards

[sidhomj]

so since the VAE is sequence level model, it outputs sequence level representations. one has to create the sample level representation by doing an average of weighted average of the sequence representations over the samples.