Add validation loss to model

poseidonchan / TAPE

Deep learning-based tissue compositions and cell-type-specific gene expression analysis with tissue-adaptive autoencoder (TAPE)

https://sctape.readthedocs.io/

GNU General Public License v3.0

48 stars 9 forks source link

Closed oyew707 closed 7 months ago

oyew707 commented 7 months ago

This change adds an optional validation loss for both reconstruction and proportions.

This change also computes the training loss for each training iteration (i.e. to match the number of epoch).