Name: Coconut (Nekopara) | AI RVC Model
Brand: Weights
Rating: 4.0 (10 reviews)

Introducing the "Coconut (Nekopara) RVC v2 Model," a masterfully crafted AI voice model, trained meticulously through 475 epochs for unparalleled accuracy and sound quality. Evolved from the Ilaria RVC mainline, this model is the epitome of precision, utilizing 13:30 minutes of high-fidelity voice data from the beloved Nekopara Vol. 4, characterized by the distinctive voice of Tezuka Ryouko. Our model is honed via advanced pitch extraction techniques, ensuring every note and nuance is captured, replicated with 19.4k training steps and optimized for a seamless integration into your RVC setup.

Revolutionary in its approach, the Coconut RVC v2 Model incorporates a regular V2 pretrain regimen, fortified by a batch size of 7 to ensure a robust and natural sound. This makes it a perfect fit for creators seeking to generate AI-driven music, craft dynamic AI covers, or bring to life engaging text-to-speech applications with a unique and recognizable character voice.

Leveraging the power of cutting-edge RVC technology, Weights empowers you to transform narratives with the authentic charm of Coconut's AI persona. Immerse yourself in the future of digital voice creation with our RVC Model and embrace the potential of AI music and AI covers. Begin your journey into effortless voice conversion and explore the sonic possibilities with our free AI tools today. Create, innovate, and revolutionize with Weights.

first model trained on ilaria rvc mainline colab trained on 13:30 minutes of her speaking from nekopara vol. 4 cv. Tezuka Ryouko pitch extraction: rmvpe steps: 19.4k steps(?) batch size: 7 pretrain: regular v2 / 40k huggingface:

Coconut (Nekopara)

Coconut (Nekopara) (RVC v2) - 475 Epochs (RETRAINED)

sxndypz

first model trained on ilaria rvc mainline colab 🤯 

trained on 13:30 minutes of her speaking from nekopara vol. 4
cv. Tezuka Ryouko

pitch extraction: rmvpe
steps: 19.4k steps(?)
batch size: 7
pretrain: regular v2 / 40k

huggingface: https://huggingface.co/sxndypz/rvc-v2-models/resolve/main/coconutv2.zip

You trained this using Ilaria Mainline beta?

Leo_Frixi

I think this model is more suitable for japanese audio than english

just have to find the clean samples from him

In case of anything, you can use RX10 or BS Roformer for deleting mic bleed.

great news, i've made the sakana model.
#H2oSakana (CEO of Phase Connect) (RVC v2) - 520 Epochs

a remaster of the kita model i made back in 2023. this character is featured in the bocchi the rock anime. trained on 12 minutes of her speaking from the "bocchi the rock! re:" and "bocchi the rock! re:re:" movies. (5.1 surround + unwa mel-roformer + mel karaoke (in some parts) + dialogue isolate (in some parts), resampled to 40k) CV: Hasegawa Ikumi pitch extraction: rmvpe steps: 17.2k batch size: 7 pretrain: klm 4 hfg / 40k hf

Kita Ikuyo (ACT2) (Kessoku Band / Bocchi The Rock!)

Kita Ikuyo (ACT2) (Kessoku Band / Bocchi The Rock!) (RVC v2) - 400 Epochs

a remaster of the nijika model i made back in 2023. this character is featured in the bocchi the rock anime. trained on 10 minutes of her speaking from the "bocchi the rock! re:" movie. (5.1 surround + unwa mel-roformer, resampled to 40k) CV: Sayumi Suzushiro pitch extraction: rmvpe steps: 16.8k batch size: 7 pretrain: klm 4 hfg / 40k hf

Nijika Ijichi (ACT2) (Kessoku Band / Bocchi The Rock!)

Nijika Ijichi (ACT2) (Kessoku Band / Bocchi The Rock!) (RVC v2) - 480 Epochs

the mashumairesh gang's now complete! this character's featured in a show by rock!! mashumairesh!! anime, alongside with the stars!! anime and the fes a live game. trained on 8:24 minutes of her speaking from the first 6 episodes of the "show by rock!! mashumairesh!!" anime. (x-minus unwa mel-roformer v2 + mel karaoke (to remove some unneeded parts that the model picked up) + uvr de-reverb-de-echo + mel denoise + bs-roformer (to clean up some moments) + rx dialogue isolate + rx de-reverb (in some parts), resampled to 40k) cv. Aya Yamane pitch extraction: rmvpe steps: 17k batch size: 6 pretrain: klm 4.3 x3 / 40k huggingface:

Ruhuyu (Show By Rock!! Mashumairesh!!)

Ruhuyu (Show By Rock!! Mashumairesh!!) (RVC v2) - 500 Epochs

the plasmagica voice model pack is almost complete! (i just need to retrain chuchu and cyan in klm 4.3 then it'll be complete) this character's featured in a show by rock!! anime, alongside with the stars!! anime and the fes a live game. trained on 4:13 minutes of her speaking from the whole season of show by rock!! s2. (x-minus unwa mel-roformer v2 + mel karaoke + bs roformer (to clean up some places) + de-reverb by anvuew (to remove reverb in some places) + rx dialogue isolate, resampled to 40k) cv. Manami Numakura pitch extraction: rmvpe steps: 14.2k batch size: 6 pretrain: klm 4.3 x3 / 40k mrf-hifigan version: (coming soon) hf

Retoree (Show By Rock!! / Plasmagica)

Retoree (Show By Rock!! / Plasmagica) (RVC v2) - 750 Epochs

retrained with klm 4.3 x3. this character's featured in a show by rock!! anime, alongside with the stars!! anime and the fes a live game. trained on (almost) 5 minutes of her speaking from the whole season of show by rock!! s2. (x-minus unwa mel-roformer v2 + mel karaoke + rx dialogue isolate, resampled to 40k) cv. Ayane Sakura pitch extraction: rmvpe steps: 13.3k batch size: 6 pretrain: klm 4.3 x3 / 40k

Coconut (Nekopara)

Description

Add a comment

Saiba Momoi (Blue Archive)

SpongeBob SquarePants (Talking And Singing)

Takanashi Hoshino (from Blue Archive)

ENHYPEN Heeseung

Sunaokami Shiroko (Blue Archive)

FIXED HATSUNE MIKU V3 - VOCALOID [SCROLL DOWN FOR LINK]

Mortis [Brawl stars]

Ariana Grande AI

Hu Tao JP (Genshin Impact)

JENNIE of BLACKPINK [Strong Ver.]

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

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

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

Tendou Arisu (Blue Archive)

Surge (brawl Stars)

Villager (Minecraft)

Jungkook (BTS)

Satoru Gojo (JJK) [VA Yuichi Nakamura]

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

Sas (Counter-strike global offensive)