Name: TupperV3 | AI RVC Model
Brand: Weights
Rating: 4.4 (3 reviews)

WebKit

Experience the techno-evolution with Weights' TupperV3, a remarkable AI Voice Model arising from the skilled application of RVC (Retrieval-Based Voice Conversion). Our enhanced TupperV3 model, advanced from its prototype TupperV2, is defined by its audio excellence. The model was trained methodically from a distinctive patch note video, diversely segmented at the 4:30 mark. The extraction and processing with GDMGS resulted in 50 highly diversified files that contribute to a rich, nuanced auditory output and superior voice alteration capabilities. A 390-epoch training enriched with the power of a TITAN pretrain yields an impeccably fine-tuned model. While explicitly engineered for speech, TupperV3 showcases exceptional performance with its integrated voice changer, further accommodating a multitude of creative AI music applications. Though singing is not its forte, it's ideal for creating unique AI Covers and Text-to-Speech rendition with unparalleled efficiency. Dive into our world and create inspiring AI Covers or Text-to-Speech adaptations with our free AI tools today! Your creativity, our innovation - let's transform the world of AI Music together.

The funny VRChat man is back and better than ever (and sounding much better). TupperV2 was never released publicly, so im penning this V3. Trained off of a single patch note video, split at the 4:30 mark, UVR'ed it, then threw it into GDMGS. The result is 50 files and a much better sounding model. I wanted to say something as the demo but I couldnt think of anything so I opted with the default lol. He does give me permission to do this (I'm not sure what his stance is on it now but I hope hes still ok with it :matsuripray: ). Use pitch 0, index doesnt really matter, but if it does for you, use 0.1 or 0.2, does not do singing, but does really well with the voice changer. (100th commit yay). Demo attached.

TupperV3

TupperV3 (50 files, 390 epochs, TITAN pretrain)

kidbot

I aint even got an image for this one man what the hell- The original (if you wanna rot your brain) can be found here, however, the clips were taken from a roblox short channel [Bloopo] (I wish i was kidding): Demo audio is attached, use pitch 10 and index 0.2. This was trained locally, and i'm sure as hell not getting that time back :bjahmeltdown:

"We're not done!" Voice, 28 files, TITAN

"We're not done!" Voice, 240 epochs, 28 files, TITAN

"Heya batboy~" ~~as you can tell i am running out of ideas for models-~~ Trained locally. Used a Season 1 video, trimmed, vocal isolated, and split using Cafeteria (dataset HF space). Feel free to react to this with a skull emoji, I wont judge Works at pitch 12 for male, 0 or -1 for female. Model:

Femboy Joker [Darkseptor]

Femboy Joker [Darkseptor] (RVC v2, RMVPE using Titan, 36 files, 240 epochs)

First model trained locally on my RTX 4060, took around 2 hours. Whispering is recommended, but can do talking as well. Original video: Demo attached. Link to model: no credit is required, just enjoy :blue_thumbsup:

Bonk! ASMR

Bonk! ASMR (33 files, DMR, 320 epochs)

You are not immune from propaganda when shes around. Side effects include fever and/or chills, running nose, swirls in eyes, and a decline in gaming skills in extremely rare chances. BTW boycott the Pillsbury dough boy. Used her placid plastic duck video as training data. Use with realtime only, sample attached.

Mallory's Asylum

Mallory's Asylum (120 files, 250 epochs, RMVPE)

Hello Vanguard, it is nice to see you again Trained off the lines in the main game (cannot give because ubisoft would slit my throat) Use pitch 11-13. Works well if you already know how to do somewhat of a female robot voice (if that makes sense lol). Index doesnt matter, but Ive got best results at 0.5-0.6. Realtime only, sample attached. Model:

TupperV3

Description

Samples

Saiba Momoi (Blue Archive)

Adel imam (Egyptian Actor) عادل امام (ممثل مصري)

JP Furina / Focalors CV: Inori Minase (Genshin Impact)

Ariana Grande AI

Hu Tao JP (Genshin Impact)

JENNIE of BLACKPINK [Strong Ver.]

Saiba Momoi (Blue Archive) (VA: Tokui Sora)

Hatsune Miku

SpongeBob SquarePants (Talking And Singing)

Takanashi Hoshino (from Blue Archive)

Satoru Gojo (JJK) [VA Yuichi Nakamura]

ENHYPEN Heeseung

Sunaokami Shiroko (Blue Archive)

Goku (Mario Castañeda) (Latin American Spanish) (UPDATE)

Villager (Minecraft)

Mortis [Brawl stars]

Jungkook (BTS)

Melodie - (Brawl Stars) - (English & Korean Voices)

Tendou Arisu (Blue Archive)

Kanye West