Higurashi Model Collection (Completed)

update: added deep voice Hanyu/Yeasomul <https://huggingface.co/Greenoguy/HigurashiRVCModels> Keiichi Maebara - RVC v2 48k, 150 epochs, 1 hour of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/KeiichiMaebaraV2.zip> Rena Ryuugu - RVC v2 48k, 200 epochs, 50 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/RenaRyuuguV2.zip> Mion Sonozaki - RVC v2 48k, 200 epochs, 50 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/MionSonozakiV2.zip> Shion Sonozaki - RVC v2 48k, 200 epochs, 50 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/ShionSonozakiV2.zip> Rika Furude - (Normal Voice) RVC v2 48k, 200 epochs, 50 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/RikaFurudeV2N.zip> Rika Furude - (Deep Voice) RVC v2 48k, 160 epochs, 54 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/RikaFurudeV2D.zip> Satoko Hojo - RVC v2 48k, 200 epochs, 56 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/SatokoHojoV2.zip> Hanyu Furude (Normal Voice) - RVC v2 48k, 200 epochs, 54 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/HanyuFurudeV2.zip> Hanyu Furude (Yeasomul/Deep Voice) - RVC v2 48k (TRAINED ON RMVPE), 200 epochs, 51 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/YeasomulHanyu.zip> samples will be below

Rena samples

Mion samples

yep

currently working on shion next then will do keiichi

soon after keiichi maybe

updated to huggingface links will probably still keep the models on my drive too as an alternative

also the list of models will now be in the description on the repo but i will post an update to here if a new model is added

update: added Keiichi and Shion will post samples

Shion sample

Keiichi sample Keiichi can also growl lol but he doesn't do it all the time (the sample showcases his growls, he can sing normally too)

now next on the list for me to work on is rika and satoko and then hanyu

added Rika has both her normal voice and deep voice

yea ive made higurashi models for other ais before they seem to do a lot better on rvc tho

next up to make is satoko and hanyu and then i can work on the other characters on the list like ooishi

added Satoko

added Hanyu this time tried different training settings as a test she's a bit breathy at times and i would've made a model of her deeper voice too but it wasn't as much data so i didn't the main models are now done for now unless i decide to retrain or something will train the others on the list sometime soon

time for funni "mmmaebara-saaaan" Kuraudo Ooishi - RVC v2 48k, 90 epoch, 40 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/KuraudoOoishi.zip> actually trained to 100 epoch but i liked 90 better

the teacher that really likes curry Chie Rumiko - RVC v2 48k, 250 epoch, 23 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/ChieRumiko.zip>

has slight vocal fry on some notes sometimes Miyo Takano - RVC v2 48k, 240 epoch, 28 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/MiyoTakano.zip>

thank you so much 🙏 this song reminds me so much of Takano

4 more models left to do Jirou Tomitake - RVC v2 48k, 200 epoch, 25 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/JirouTomitake.zip>

the song choices for the samples are so good

3 more models to go now Kyousuke Irie - RVC v2 48k, 150 epoch, 28 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/KyousukeIrie.zip>

2 more models left also Satoshi works with girl inputs lol Satoshi Hojo - RVC v2 48k, 150 epoch, 22 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/SatoshiHojo.zip>

1 more model left finally lol trained to 100 epoch actually but 90 was better also sings aggressively sometimes like Keiichi Mamoru Akasaka - RVC v2 48k, 90 epoch, 36 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/MamoruAkasaka.zip>

and with this all the models are completed (unless i decide to retrain or something lol) his voice is pretty rough Tatsuyoshi Kasai - RVC v2 48k, 150 epoch, 22 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/TatsuyoshiKasai.zip>

Higurashi Model Collection (RVC v2) (Completed)

quick link back to the main models (also in total: 16 models, if counting rika's deep voice as an extra model, otherwise 15 models)

didn't really extract it myself it was all from the 07th mod since they had the raw voice files (except rena, mion, and shion their files were from some crossover game that were true 48khz) but since i found new longer data from 07th mod that is true 48khz i might update the models

dunno haven't seen that file type before

updated the first batch of characters with retrained models on newer dataset so now the samples above are outdated they should sound a bit more clearer and have more data

Jirou Tomitake - RVC v2 48k, 200 epochs, 46 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/JirouTomitakeV2.zip> Miyo Takano - RVC v2 48k, 200 epochs, 52 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/MiyoTakanoV2.zip> Kyousuke Irie - RVC v2 48k (Trained on RMVPE), 150 epochs, 56 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/KyousukeIrieV2.zip> Chie Rumiko - RVC v2 48k, 200 epochs, 27 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/ChieRumikoV2.zip> Kuraudo Ooishi - RVC v2 48k, 200 epochs, 51 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/KuraudoOoishiV2.zip> Satoshi Hojo - RVC v2 48k, 300 epochs, 19 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/SatoshiHojoV2.zip> Mamoru Akasaka - RVC v2 48k (Trained on RMVPE), 200 epochs, 53 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/MamoruAkasakaV2.zip> Tatsuyoshi Kasai - RVC v2 48k (Trained on RMVPE), 300 epochs, 18 minutes of data <https://huggingface.co/Greenoguy/HigurashiRVCModels/resolve/main/Version2/TatsuyoshiKasaiV2.zip> updated second batch of characters to newer dataset (no new samples rn but they should be improved, also testing rmvpe training on some models) now ill work on hanyu's deep voice and some extra characters

added yeasomul/deep voice hanyu to main post sample

i want to download rika furude's deep voice, please owner :3

and her normal voice too