Trained on 22/24 minutes of data from the ElevenLabs site (rx11 de-ess + de-click). This model was trained with dataset of pure talking, so it's better to use it for talking, but it can sing too.
Pitch extraction: RMVPE
Steps: 13.5k
Batch size: 16
Pretrain: Original v2 / 32k
Precision: FP32 (?)
This was trained on RVC Mainline (Lightning.ai) with A10G GPU, learned how to use it thanks to Litsa :p
Don't forget to credit my YT channel if used: @fitzyrvc2024; i would appreciate it.
Huggingface: https://huggingface.co/FurnTheFurnace/RVCModels/resolve/main/Adam-Elevenlabs-Spanish_V2_e300_s13500.zip?download=true
Weights.gg: https://www.weights.gg/models/cm06spua700713fuxuxjyk347