Description

This model was made using RVC V2 (mangio-crepe). The dataset that I used came up to around 15 minutes worth of audio, other weights are inside of the rar as well (Because I have no clue how to open tensorboard so I just wing it). If you're curious about the different versions, the one which is normal had all the UVR models I typically use put onto the dataset while the reverb version one only had one UVR model put onto the dataset which retains some of the reverb but less buzz perhaps? Up to you to find out which one you prefer. Side Note: Getting the dataset was a gigantic pain in me arse and its kind of "iffy" but if anyone wants to try their hand on perhaps finding ways to improve it I keep all the files from my models so I can send it over if you want to see if you can make it better in any way! Model By: Me Training By: Me again (NORMAL VERSION No Reverb Most UVR Models Applied to Dataset) Link to Model: (REVERB VERSION Only One UVR Model Applied) Link to model: