Name: Sara Hoshikawa (Nijisanji) | AI RVC Model
Brand: Weights
Rating: 4.4 (12 reviews)

WebKit

 Meet Sara Hoshikawa, our advanced AI Voice Model trained using the latest Retrieval-Based Voice Conversion version 2 (RVC v2), along with cutting-edge techniques such as RMVPE and 330 epochs. Sara offers remarkably humanlike speech output, perfect for text-to-speech applications. Aside from speaking, she excels at creating AI covers of popular songs by converting existing audio into her unique voice. Ideally suited for musicians, podcasters, animators, and hobbyists alike, Sara brings artificial intelligence to new heights through seamless integration with RVC technology. Discover how you too can generate AI covers or utilize our top-tier text-to-speech capabilities absolutely free with our innovative AI tools at Weights. Explore limitlessly, as your creativity meets its match with Sara Hoshikawa's exceptional performance.

length: 52 mins (talking and singing) Trained using: latest original RVC with some code tweaks

Sara Hoshikawa (Nijisanji)

Sara Hoshikawa (Nijisanji) | RVC v2, RMVPE, 330 Epochs

makidanye

using: Applio (main branch post 3.2.8-bugfix) Dataset length: 6 mins (talking) Pretrain, vocoder & sample rate: DMR v2, default hifigan, 32k F0 method: crepe (32 hop length) Embedder: contentvec (default) Batch size: 8 Training precision: fp32 Note: you can try other model checkpoints posted in the comment below.

Usada Pekora (Hololive), DMR

Usada Pekora (Hololive) | RVC v2, 202 Epochs, DMR v2

using: Applio 3.2.8 bugfix (no new vocoder) Dataset length: 20.5 minutes Pretrain & sample rate: KLM4.3x2, 32k F0 method: rmvpe Embedder: contentvec (default) Batch size: 8 Training precision: fp16 (default)

Amity Blight (The Owl House)

Amity Blight (The Owl House) | RVC v2, 219 Epochs

using: Hina's mainline Kaggle notebook Dataset length: 33.5 mins Pretrain & sample rate: Default, 32k F0 method: rmvpe Batch size: 8 Training precision: fp16 (default)

Kureiji Ollie (Hololive ID)

Kureiji Ollie (Hololive ID) | RVC v2, 273 Epochs

using: Hina's mainline Kaggle notebook Dataset length: 33.5 mins Pretrain & sample rate: Default, 40k F0 method: rmvpe Batch size: 8 Training precision: fp16 (default) Note: Has anime adaptation in summer 2024 but the dataset processed from the visual novel voicelines.

Minamo (ATRI -My Dear Moments

Minamo (ATRI -My Dear Moments-) | RVC v2, 183 Epochs

using: Hina's mainline Kaggle notebook Dataset length: 2 hours Pretrain & sample rate: Default, 40k F0 method: rmvpe Batch size: 8 Training precision: fp16 (default) Note: Has anime adaptation in summer 2024 but the dataset processed from the visual novel voicelines.

Sara Hoshikawa (Nijisanji)

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