Aloy (HFW) - (32 Hop Length) - 32k sample rate - 8 batch size

Description

Model created by me (AI_Characters). Please credit me for the model creation when uploading a piece of work (song, speech, etc...) you created using this model to social media (YouTube, Instagram, etc...). I would love to know when it is used! If you want to support what I am doing, donations to my Ko-Fi (< would be greatly appreciated! Those will go towards funding the renting of GPUs for further experimentation and model making. < Dataset Source: YouTube video of 4h of ingame voice lines of Aloy from Horizon: Forbidden West - Silence truncated - Final length: 10min Training settings Model: RVC-v2 - Pitch extraction algorithm: Mangio-Crepe with 32 Hop Length - Target sample rate: 32k - Batch size: 8 - Epochs: 79 Recommended inference settings Pitch extraction algorithm: RMVPE - Median filtering: 3 - Search feature ratio: 0-0.5 *(higher than 0 will create a more true-to-the-character voice but also introduces more artifacts)* - Volume envelope: 0-1 *(highly depends on the song)* - Protect voiceless consonants: 0.1 Extra notes: Unfortunately due to the nature of the source dataset, the model sounds breathy. I will see if I can source a better dataset where Aloy does not sound constantly out of breath. *Any honest critique of the model is greatly appreciated! Please post such in this thread! Please also post any works you created using this model in this thread, too!*