Name: Jelly Hoshiumi (Phase Connect / VTuber) | AI RVC Model
Brand: Weights
Rating: 4.8 (7 reviews)

WebKit

Unveil the true essence of virtual euphony with the "Jelly Hoshiumi (Phase Connect / VTuber) (RVC v2) - 380 Epochs" AI Voice Model, a groundbreaking innovation from Weights that stands at the pinnacle of RVC Model technology. Crafted with precision from approximately 16 minutes of premium-quality audio harnessed from her Valentine's Day stream, this RVC Model represents the cutting edge in AI Music and voice synthesis. The Jelly Hoshiumi AI Voice Model has been meticulously trained through 380 epochs, ensuring the voice retains not only the pitch and tone but also the unique nuances of this immensely popular VTuber personality.

Perfect for content creators and fans who yearn to create AI Covers or pilot their very own Text-to-Speech projects with unprecedented realism, this model exemplifies the seamless blend of art and artificial intelligence. Invite Jelly Hoshiumi's digital voice to elevate your creative endeavors, from YouTube parodies to podcast intros. With Weights' RVC v2 technology, you can effortlessly emulate Jelly Hoshiumi's signature sound, endowing your content with a captivating virtual presence that engages audiences.

Experience the next generation of virtual voice technology with Weights. Create AI Covers or innovate your Text-to-Speech applications with our free AI tools today, and let the symphony of the future commence.

trained on (almost) 16 minutes of data from her valentine's day stream (melband roformer + rx10 de-reverb) pitch extraction: rmvpe steps: 19.3k batch size: 7 pretrain: regular v2 huggingface:

Jelly Hoshiumi (Phase Connect / VTuber)

Jelly Hoshiumi (Phase Connect / VTuber) (RVC v2) - 380 Epochs

sxndypz

now i realize i had de-verb a little too high

eh, i'll redo the dataset and retrain it later

https://tenor.com/view/panko-komachi-panko-slurs-vtuber-phase-connect-gif-1869457431532208483

yay an another phase connect model after months trained on 17 minutes of audio data from the rating viewers' desks stream (mel-roformer fv7z + eq + de-plosive + de-click + de-ess, resampled to 32k) pitch extraction: rmvpe steps: 18.8k batch size: 7 pretrain: klm 4 hfg / 32k huggingface:

Mari Mari / marimari_EN (Phase Connect)

Mari Mari / marimari_EN (Phase Connect) (RVC v2) - 320 Epochs

initially i was going for 475e but then it sounded awful and the 350e version sounded a bit more stable so i went with that. it's also a test to see how the model trained on klm pretrain would handle lossy audio as dataset this character's featured in blue archive. trained on 11 minutes of audio data from both jp and global versions since both of them have different lines. (eq + spectral de-noise) CV. Tanezaki Atsumi pitch extraction: mangio-crepe / (applio) crepe - 64 hops steps: 14k batch size: 7 pretrain: klm 4 hfg / 40k hf

Yurizono Seia (Blue Archive)

Yurizono Seia (Blue Archive) (RVC v2) - 350 Epochs

the fraise model also hasn't been remastered since rvc v1 was around, so i remastered that too. this character's featured on most nekopara series. trained on 15 minutes of audio data extracted from nekopara after. (slight de-noise) CV: Nozomi Yamamoto pitch extraction: rmvpe steps: 18.9k batch size: 6 pretrain: klm 4 hfg / 40k hf

Fraise (ACT2) (Nekopara)

Fraise (ACT2) (Nekopara) (RVC v2) - 300 Epochs

this model hasn't been remastered since rvc v1 was around, so i remastered that. this character's featured on most nekopara series. trained on 9 minutes of audio data extracted from nekopara after. CV: Inokuchi Yuka pitch extraction: rmvpe steps: 18k batch size: 6 pretrain: klm 4 hfg / 40k hf

Vanilla (ACT2) (Nekopara)

Vanilla (ACT2) (Nekopara) (RVC v2) - 500 Epochs

i finally remastered a chocola model after 2 years. happy now? this character's featured on most nekopara series. trained on 12 minutes of audio data extracted from nekopara after. CV: Himari (aka Tomonaga Maki) pitch extraction: rmvpe steps: 18k batch size: 7 pretrain: klm 4 hfg / 40k hf

Jelly Hoshiumi (Phase Connect / VTuber)

Description

Samples

Gorillaz - Noodle [PHASE 2]

Shylily

Bao (indie vtuber)

Mashiro Kanon (Vtuber/streamer)

Silvervale (Former VShojo VTuber, now independent)

*namirin ("In hiatus" VTuber, famous for Renai Circulation cover)

Girl_DM_

Yuko Yurei - Vtuber

Obake Pam EN VTuber

2D (Phase 1 Gorillaz)

StarVtuber (Vtuber - Spanish Streamer/Youtube)

SuRge

Koa Akugaki Vtuber (es/en/jp)

Dyarikku VTUBER TWITCH STREAMER

RakunVT

Kizuna AI Vtuber

Sinder

Nihmune Trained with 30min on clear audio

Linh Lan Ch. (Vtuber Vietnam)

Momosuzu Nene (Hololive jp Gen5 / Vtuber)

Nimu (Latin Vtuber)

Hibiki Du Ca

Akatsuki Ban Mai (NIJIGEN VTUBER Gen1)

Clone Trooper Phase 2

shxtou

Maemi Tenma

Dacapo

Shxtou

綾音Aya (Chinese VTuber)

Pecorine (Princess Connect Re:Dive)

Meica / Hina Misora - Spanish VTuber

Kirche

Veibae (VTuber, EN)

BuffPup Vtuber

Riifu Evergreen (vtuber, ex-EIEN Project)

Rainhoe 32k Sample

Alec Watson / Technology Connections

Fallenshadow / shondo Vtuber

Peanut Butter Jelly Time

Thien Tra Ch (NIJIGEN VTUBER)

Ha Chi Ch. (VTUBER NIJIGEN)

"Connection Terminated" Speech (From FNAF 6)

Shigure Ui (Illustrator, Vtuber)

Shylily

Mana Renewal (Ru Vtuber)

Uki Violeta

2-D (Phase 1 Gorillaz)

Noriaki Sugiyama (Actors Song Connection Era)

Emikukis

Meica / Hina Misora - Spanish VTuber