Name: balbal(eko) (Infamous AI Hubber) | AI RVC Model
Brand: Weights
Rating: 4.0 (3 reviews)

WebKit

Introducing balbal(eko), our cutting-edge AI Voice Model from Weights, designed exclusively for the innovative RVC (Retrieval-Based Voice Conversion) technology. This model, also known as the Infamous AI Hubber, boasts an impressive repertoire tailored predominantly for AI music and creation of captivating AI covers. Embellished with a unique dataset largely comprised of melodious singing voice samples, the balbal(eko) model guarantees remarkable performance for all your auditory projects.

Equipped with the RMVPE architecture and fine-tuned over 37,000 steps with a 32k sampling rate, the balbal(eko) RVC model thrives at maintaining vocal integrity. Benefiting from a meticulous retraining process on a specialized dataset provided by @Barkert, this model ensures your AI-generated covers exude a rich and authentic sound that resonates with your audience. And with 3300 epochs expertly calibrated to avoid overtraining, the balbal(eko) can deliver a consistent, high-quality voice output for your creative needs.

Step into the future of AI-generated music with Weights' balbal(eko) RVC Model. Unlock the potential of voice conversion and let your imagination soar as you create AI covers with our free AI tools. Begin crafting your unique soundscapes today by taking advantage of the state-of-the-art capabilities at your fingertips — your next AI-powered vocal masterpiece awaits. Explore the possibilities and transform your text-to-speech projects with Weights' leading technology.

Retrain of @Barkert - ()'s dataset he sent me. Dataset is composed of mostly singing. It has alright performance with singing. Haven't tried it for realtime ngl. 5 minute dataset. I promise it's not overtrained; 3300 epoch

balbal(eko) (Infamous AI Hubber)

balbal(eko) (Infamous AI Hubber) (RMVPE, 32k, Ov2, 37k steps)

unitedshoes

Retrain of @Barkert - バルバル(エコ)'s dataset he sent me. 

Dataset is composed of mostly singing. It has alright performance with singing. Haven't tried it for realtime ngl. 5 minute dataset. I promise it's not overtrained; 3300 epoch

Download: https://huggingface.co/UnitedShoes/balbal1/resolve/main/balbal1_3375e_37125s.zip?download=true

realpikachuwu

sounds like shit on anything higher pitched btw

Were talking about the Wii today so lets get this out of the way (Proceeds to fastball controller at tv) Trained on 8 mins 20 seconds of singing data, sourced from Scott's many musical skits, compiled here: Some songs and pieces were removed for quality/artifacting. Downloaded from YouTube full quality, split with UVR, cleaned with RX11 and prepared with Audacity. Despite my best effort, model is slightly noisy, but fuck it deal with it its a youtube>uvr voice model Model should perform well on most singing vocals. If you need a speaking model, there's other great models already posted. F0: RMVPE Length: 8:20 Sample rate: 32k Pretrain: OG v2 32k RVC client: Applio 3.2.8 Batch: 12 Epochs: 700 Steps: 14,000

Scott The Woz

Scott The Woz (Singing, RMVPE, 700e)

Supreme AI Hub brainrot Trained on the full vocals from Ballin' by Mustard (feat. Roddy Ricch). Used BS Roformer for the initial split from the FLAC, then used RX 10 to clean the audio and remove reverb. Model works great with ballin

Ballin (Vocals)

Ballin (Vocals) (RMVPE, 4625e, 37,000 steps)

Trained on studio session files from Ye era (2017-2018) Model

Kanye West

Kanye West (Ye Era, RMVPE, 1900e, 24700 Steps)

Legendary computer virus guy RMVPE, 1000 epoch (17000 steps, began overfitting) 32k, Ov2-Super. Dataset was ~10 minutes long Model

balbal(eko) (Infamous AI Hubber)

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