Open ksasso1028 opened 7 months ago
Thank you for your comment, we haven't compared it to L1 loss, but we are thinking of adding it.
awesome, certainly would be interested across the board on a few different datasets. Planning to experiment with it in the context of audio (spectrograms or raw waveforms) and compare to L1
great work, in the context of reconstruction have you compared RLP to l1 loss? Did not see mention of it in the paper.