Description

SalC1 RVCv2 300 Epochs (Ultra Quality Raw Audio Dataset) : 2 Hours / 120 Minutes Algorithm: RVMPE Quality: 40k This RVCv2 model is quite possibly the finest in terms of quality. It was trained using 2 hours of impeccable American speech, employing RVMPE pitch extraction and epoch 300. Through persistent efforts, i managed to minimize the mel/loss and g/loss to their lowest achievable levels.