Name: Ozen (Made in Abyss) | AI RVC Model
Brand: Weights
Rating: 4.0 (11 reviews)

WebKit

Introducing our RVC Model called Ozen from Made in Abyss (175 epochs, KLM 4). This exceptional AI Voice Model has been meticulously trained on a 6-minute dataset using 175 epochs with a batch size of 7, demonstrating Weights' capability in crafting top-tier RVC Models for seamless AI Music generation. Ozen's uniqueness stems from the incorporation of the KLM 4 pretrain, a distinguished strategy used to optimize the learning curve and elevate the overall performance of our model. With a superior 40k sample rate, this model assures the highest quality output that can cater to any of your AI Music requirements. The language of Ozen is Japanese, making it an optimal choice for those seeking to create AI Covers or Text-to-Speech in Japanese with unlimited creativity. Experience the Weights difference and explore the boundless possibilities in AI Music. Encapsulate your thoughts and styles through Ozen. Dive into the world of AI Music by creating your AI Covers or Text-to-Speech today, using our free AI tools available at Weights.

This model was trained on 6 minute dataset with KLM 4 pretrain, 175 epochs, batch size on 7, 40k sample rate

Ozen (Made in Abyss)

Ozen (Made in Abyss) (175 epochs, KLM 4)

Joxi

This is remake of my old model. This model was trained on 4:28 length dataset with KLM 4 pretrain, 300 epochs (8.7k steps), batch size on 4, 32k sample rate

Rudeus Greyrat (Mushoku Tensei)

Rudeus Greyrat (Mushoku Tensei) (KLM 4, 300 epochs)

This model was trained on 3:13 length dataset with KLM 4 pretrain, 300 epochs (6k steps), batch size on 4, 32k sample rate

Takopi (Takopi's Original Sin)

Takopi (Takopi's Original Sin) (300 epochs, KLM 4)

This model was trained on 2:41 length dataset with KLM 4 pretrain, 300 epochs (5.4k steps), batch size on 4, 32k sample rate.

donk (CS 2 / Team Spirit)

This model was trained on 4:39 length dataset with KLM 4 pretrain, 300 epochs (9k steps), batch size on 4, 32k sample rate.

karrigan (CS 2 / FaZe Clan)

This is remake of my old model. This model was trained on 5:24 length dataset with KLM 4 pretrain, 220 epochs (8.14k steps), batch size on 4, 32k sample rate.

Ozen (Made in Abyss)

Description

Samples

How It's Made voice

Mococo Abyssgard (FUWAMOCO) (Hololive EN gen 3)

SZA (Made by Tyraskin)

Vueko (Made in Abyss)

Faputa (Made in Abyss)

Madeline Hatter (Alondra Hidalgo) (Ever After High) Español Latino

Nanachi (Made in Abyss)

madeleine cookie

Tord(Beyond/un-official, 7 tord fan made voices mixed)

baldloon from attack the baldloons. (I had an old one but I made it better and hopefully it comes faster)

Prushka (Made in Abyss)

made this to goof off

Abyss cookie

Madeline from celeste

Lomando.com Groan (basically the jumpscare that made mutahar laugh)

Madea

Faputa (Made in Abyss)

Vegeta (Orginally made by @nicegame)

Panic! At the Disco - Hey Look Ma, I Made It (Drums)

How It's Made Narrator Girl [Castillian Spanish]

MADEN BY EVAR

Madeline Tracey Lee Smythe

Madeirol

Bondrewd(Made in abyss)

Sister Madeline/Evil Nun 1

Madeleine McGraw

Sugar Dimples Madeline

Nanachi [Made in Abyss]

Tori(Unofficial/5 diiferent fan made voices mixed into one)

Clubbox - Fan-Made My Singing Monsters Characters

A random voice I made due to boredom

Freestyle 3 Ken Carson Voice (made by YvlLocky)

Singer from All The Single Furries, the worst model i've ever made

Fuwawa Abyssgard (FUWAMOCO) (Hololive EN gen 3)

Understandable Pyro (TF2 Mod Made by FlowerX)

Brooks Moore (How Its Made)

Riko (Made in Abyss)

Madeline (Celeste)

How It's Made Narrator Main Narrator (Castillian Spanish)

Mococo Abyssgard (FUWAMOCO/Hololive Advent)

Jaira Burns (Made Here)

Politea (Winx Club: Mystery Of The Abyss)

Rafayel Abysswalker

Don’t use this shi I made for my friends and I hate him so use the other Bree rabbit instead

Paxton(Eddsworld Patryck Fan made Neighboor)

JOSE MADERO

Reg (Made in Abyss)

Quoin God (A fan made Super Mario character)

Abysstale Sans

Lemonnyy (Made By Lemonn)