Question: Get the loss from the model

Hi,
Anybody who knows how to get the loss from the model? Like getting rmse from some regression package - I'd like to get warp, warp-os losses for parameter tuning, finding number of epochs and similar. I don't think it is available, but maybe I am wrong. Or maybe someone has some implementation that they want to share?
Best regards and thanks in advance

Start by reviewing LightFM's model and training API, focusing on how WARP and WARP-OS losses are represented. No files or tests are named; first clarify whether the desired outcome is documented loss access or a new API, then define completion as a supported way to retrieve those losses for tuning epochs and parameters.

Written by the indexing model from the issue text.