Name: Koseki Bijou | AI RVC Model
Brand: Weights
Rating: 4.4 (22 reviews)

WebKit

 Introducing the Koseki Bijou [RVC v2] [RMVPE] AI Voice Model, specifically designed for use with Retrieval-Based Voice Conversion (RVC) technology. This advanced model has undergone rigorous training through 635 epochs and 83,820 steps, resulting in highly accurate voice simulations. Derived from HoloAdvent's Hololive English series, this model captures the essence of Bijou, allowing users to generate realistic speech patterns and intonations similar to those heard in over 52 minutes of her dialogue. With applications ranging from creating AI covers to generating text-to-speech content, this powerful tool offers limitless possibilities for engaging your audience. Start exploring these innovative capabilities today by visiting Weights' user-friendly platform and unlocking access to top-tier AI music and audio solutions – all completely free!

Origin: HoloAdvent (Hololive English) minutes and 52 seconds of Bijou talking Batch size per GPU used during training: 8 Image made courtesy of Mo Kitada Credit would be much appreciated if you used this model Discord: or YouTube: @AIDontEvenKnow-cd8qe You are solely responsible with what you create with this model so use wisely.

Koseki Bijou

Koseki Bijou [RVC v2] [RMVPE] 635 Epochs 83820 Steps

AIDontEvenKnow

This audio sample was inferred on the Aesthetic RVC Inferrence Hugging Face Space by r3gm using the RMVPE+ pitch extraction algorithm with "Split Audio" enabled on a minimum value of 13 Hz and a maximum value of 350 Hz.

myura1983

Origin: Rock music Trained on: 48 seconds of Meissner singing: 26 seconds from *Dangerous Games* 22 seconds from *Windows to Light* Save frequency: 10 epochs Log interval: 100 steps Batch size used during training: 3 Sample rate: 32k Image made courtesy of unknown photographer Credit would be much appreciated if you used this model Discord: @AIDontEvenKnow or YouTube: @AIDontEvenKnow-cd8qe You are solely responsible with what you create with this model so use wisely.

Stan Meissner

Stan Meissner [RVC v2] [RMVPE] 210 Epochs 2100 Steps

Origin: Progressive rock music (most notably: King Crimson, U.K. and Asia) Epochs: 670 Steps: 6700 Trained on: 02 minutes and 32 seconds of Wetton singing songs by King Crimson and U.K. from live recordings: * 17 seconds from Starless and Bible Black by King Crimson * 49 seconds from USA by King Crimson * 01 minute and 24 seconds from Night After Night by U.K. Save frequency: 10 epochs Log interval: 100 steps Batch size used during training: 10 Sample rate: 32k Image made courtesy of Mike Inns Credit would be much appreciated if you used this model Discord: @AIDontEvenKnow or YouTube: @AIDontEvenKnow-cd8qe You are solely responsible with what you create with this model so use wisely.

John Wetton

Origin: HoloMyth (Hololive English) Epochs: 455 Steps: 15925 Trained on: 59 minutes and 55 seconds of Kiara talking Save frequency: 5 epochs Log interval: 175 steps Batch size used during training: 32 Sample rate: 32k Image made courtesy of Kurohoshi (@96hoshii_0317) Credit would be much appreciated if you used this model Discord: @RigidSpinner or YouTube: @AIDontEvenKnow-cd8qe You are solely responsible with what you create with this model so use wisely.

Kiara Takanashi

Origin: Whitesnake Epochs: 950 Steps: 7600 Trained on: 01 minute and 53 seconds of Coverdale singing songs by Whitesnake: * 59 seconds from Slide It In * 31 seconds from Whitesnake * 22 seconds from Slip of the Tongue (Note: The recordings used are the 2020 remixes derived from Love Songs the Rock Album and the Blues Album) Save frequency: 25 epochs Log interval: 200 steps Batch size used during training: 4 Sample rate: 32k Image made courtesy of Metal Magazine (Argentina) Credit would be much appreciated if you used this model Discord: @RigidSpinner or YouTube: @AIDontEvenKnow-cd8qe You are solely responsible with what you create with this model so use wisely.

David Coverdale [1983-1989 Era]

Origin: HoloMyth (Hololive English) minutes and 59 seconds of Mori talking Batch size per GPU used during training: 4 Sample rate: 32k Image made courtesy of Kurohoshi (@96hoshii_0317) Credit would be much appreciated if you used this model Discord: @RigidSpinner or YouTube: @AIDontEvenKnow-cd8qe You are solely responsible with what you create with this model so use wisely.

Koseki Bijou

Description

Samples

Koseki Bijou

Bijou (Italian Dub, Giovanna Papandrea)

Koseki Reina

Koseki Bijou

Ribon-chan (Japanese Bijou, Kazusa Murai)

Bijou (Japanese Dub, Hamtaro)

Hamtaro - Bijou

Colitas (Latin Spanish Bijou, Leyla Rangel)

Bijou (Ribon-Chan) (Hamtaro Series)