Ilmu yang bermanfaat bang. ini baru praktek ternyata bisa juga akhirnya. sisanya paling pinter2 ngeracik biar suara AI nya nggak fals. trims tutornya bang

Ga nyangka yang bikin orang indo :V Terakhir pakai sekitar tahun lalu dari tutornya bule tapi tiba tiba layoutnya berubah semua, dan akhirnya github membawaku kemari

akhirnya, thank you tutorianya. udah 2 bulan mangkrak channel ku 😭

Also, once you fixed that Cuda problem, can you please also add a option to leave the songs unstemmed and instead just goes straight into the inference? It will be great for long voice audios with very little to no background noise

Bang mau tanya cara ngatasin runtime error/filenotfound tu gimana bg??

makasih bang... It worked!!

Colab AICoverGen yg Hina stuck di misahin vocal sama instrumentalnya bang muncul tulisan "Failed to create CUDAExecutionProvider" di outputnya

Wahhhh makasih tutorialnya bg😁

While the inference has problems, i tried using the voice changer for conversion. But it requires a good internet connection

I deeply thank you for giving this tutorial. But recently, i've noticed a problem. The WebUI simply wouldn't load, and gives a 504 gateway timeout. Any ways to fix this problem?

Naisu

bg ini prosesnya lebih cepat kalo pkai runxtime cuda 12, tapi error hasilnya di output? kenapa bsa begitu bang?

The Cuda problem is not a Gradio ui problem, because if that was the case, itwould've been automatically disconnected.

mantep

Makasih bang, lu kerenw

Izin bertanya bang, apakah ada cara buat bikin ini semua jadi offline? Biar kalau google ngegocek lagi kita masih bisa training di offline

kak kok stuck di separating vocals from instrumental yah😢

Failed to create Cuda execution provider problem please fix i really this

Think you

The gradio link won't show up for me :(

kok saya dibagian rvc model dibagian download model resources pdhl saya udh aktifin akses gdrivenya kok masi private?

ada voice changer terbaru tanpa google collab yg pro gak bg?

bang kalau eror no cuda gpu are available cara ngatasinnya gimana?

Ternyata make model di huggingface gabisa ya, udah ane coba download model zipnya tuh gabisa, hasilnya cuma create folder aja

Damn aku coba pake android tetep bisa bang, terimakasih tutorial nya.

Bang.. tutor yg aku req bot Enka Network itu blom jadi kah? 😢

apakah ini harus diupdatekah bang,soalnya? soalnya pas masuk ke web gradio live ga bisa masuk masuk

Bang bagaimana kalau maunya cuma vocalnya doang?

index -1 is out of bounds for axis 0 with size 0 , maksud nya , gw dh ngabisin 1jam , lebih enakkan ai versi lama drpd yg baru ribet

Sekarang udah ga bisa ya bang? Soalnya ga ada pencet runtime nya

Stuck di clone repository bg?

*No CUDA GPUs are available* so sad

Hello Ryusui! i have some questions, i used AICoverGen WebUI and i used the Gardio link to make ai song covers! but i only could do it in a short time, because it then had a error everytime i wanted to make a new cover! do you know what i did wrong?

@ryusuikagaku

7 ай бұрын

which google colab do you use

@dracovishfan2023

7 ай бұрын

@@ryusuikagaku i used the AICoverGen Mod by Me & Hina (Colab Free)!

@ryusuikagaku

7 ай бұрын

hmm, i just do 12 cover yesterday and not getting any error with that colab@@dracovishfan2023

@dracovishfan2023

7 ай бұрын

@@ryusuikagaku that's weird... 🤔

@thanh2425

7 ай бұрын

What error is it

hina collabs gak bisa lg yah kak

nah gradio yg ini yg paling bagus hasilnya nih

Inference nya sekaligus sama instrumental ya bang, bisa gk generate nya cuma vokal doang, kadang kalo sama instrumen nya malah ikut ke cover juga

@ryusuikagaku

8 ай бұрын

Programnya otomatis misah vokal dan instrumen, dan cuma inference vokal

Ok might be the last report here but it seems like the cuda problem still persists. While the RVC AIcovergen is having problems, I advise you to use use the real time voice changer as an alternative, although you do need a great internet, a moderately strong device as well, if you're doing audio capture. And finally, if you're doing song covers you have to stem it manually using a different program.

@NotATaxiDriver

7 ай бұрын

I found out how to fix it, it's caused by the Cuda version being newer than currently supported by onnxruntime

@thanh2425

7 ай бұрын

@@NotATaxiDriver that's just where the problem came from I'm looking for a method to actually fix it

@NotATaxiDriver

7 ай бұрын

I did fix it, I forked the colab link and edited it. Unfortunately, youtube won't let me comment the link @@thanh2425

@NotATaxiDriver

7 ай бұрын

@@thanh2425 colab(dot)research(dot)google(dot)com/drive/1khoMO4_xHCqkf0AjwkW71NY8uJ7-NCw_?usp=sharing there, just replace (dot)

@thanh2425

7 ай бұрын

​@@NotATaxiDriverIt just redirects to google drive, which will display a error message. Can you try another way

bang izin nanya, kenapa pas saat Run WEB UI malah ga muncul link untuk akses ke 😃dalam web ai nya?

@Ritsu_subhi

5 ай бұрын

Sama bang, baru error hari ini, biasanya lancar lancar aja

kak, kok publik link nya 505 error ya?

min kalau mau masukan sampel kita sendiri bagai mana min..? saya putar videonya sampai habis ngak ada di jelasakan. .. mohon infonya min..

@ryusuikagaku

7 ай бұрын

sampel sendiri tu gimana? suara sendiri? ada di video sebelum ini

Semoga mbah google kali ini kayak player epic, buta map 🙏😣

Bang bantu Masalah Error Cuda

bang kalo rvc mangio bisa kelabuhi juga ga? kek ai cover ini, kalo bisa dibuatin dong

@user-dc6fv1gk9q

8 ай бұрын

ini kan yg versi ai cover kalo yg rvc mangioo

bang mau nanya, kenapa bisa ada onxruntime cuda 12, yg dimana loadingnya lama? di google colab? alasannya apa? itu buat ngeupdate sistemnya atau apaan ya bg?

@ryusuikagaku

7 ай бұрын

ngedowngrade versi CUDA di google colab. kalo ga di downgrade, error

@saofans

7 ай бұрын

thank you@@ryusuikagaku

cara bikin link hunggingface gimana bang

When are you going to fix the cuda 12 problem? or are you're working on it now?

@ryusuikagaku

7 ай бұрын

i already add cuda downgrade code to the colab

@thanh2425

7 ай бұрын

@@ryusuikagaku okay thank you. Will try it later

@thanh2425

7 ай бұрын

​​@@ryusuikagakuIt worked. But now there's only one problem left to solve: and that is the inability to upload models. Everytime i did it it just does nothing for a very long time, and after a while it creates an empty folder, then displays error.

bang kok gw di step pertama ga ada pop up connect to google drive?

@thanh2425

7 ай бұрын

Try running it again

bang tutorial bikin model suara sendiri di video mana bang? viewer baru nih

@ryusuikagaku

8 ай бұрын

Video sblm ini

Udah kayak cracker cuy /dikira maling kita gegara pakai collab gratisan doang.. Susah di rekomenin untuk pengguna tetap, tiap waktu slalu di ban google😂..

When is it going to support Google Drive download? I tried uploading the models to pixeldrain myself because i don't have the official link and the link doesn't work. Just show error. Link type is the same. I have to upload the models stored in my devices instead cause that's the only option. Also thank you very much for the tutorial. Now i can make Rick Astley sing Tokyo Fantasy in under 2 minutes :) the other method in huggingface took too long and it's unstable.

@ryusuikagaku

8 ай бұрын

It already suport it. Watch 05:40

@thanh2425

8 ай бұрын

​@@ryusuikagakuok the google drive download actually work, which is not the case the first time i ran it. Maybe slow internet is to blame. Anyways i wonder how did they even make the gui accesible for free users?

@ryusuikagaku

8 ай бұрын

@@thanh2425 they encode some code inside the colab so google will be hard to detect it. I explain it also on the video

Bang intro mu itu gimana cara buatnya, mangap"nya bisa sesuai? 😅

@ryusuikagaku

7 ай бұрын

Ngepasin scene manual, ga ada editan tambahan wkwk

Aku butuh nama channel ardha aja, lebih ikonik soalnya

@ryusuikagaku

8 ай бұрын

Ni lebih 😎

Kemarin2 bisa, skrg mentok di downloading song, ga jalan ttep di 0%, kenapa ya bang

@ryusuikagaku

8 ай бұрын

Refresh atau jalanin ulang

Bang punya saya kalo index ratenya diatas 0 ada error "index -1 is out of bounds for axis O with size 0" itu gimana ya cara ngatasinnya? Soalnya kurang berasa covernya kalo indexnya 0

@ryusuikagaku

8 ай бұрын

Udah yakin di dlm zip modelnya ada file index?

@Ritsu_subhi

8 ай бұрын

​@@ryusuikagakudi folder Zip ada file "added_IVF1958_Flat_nprobe_1_Itsuki_v2.index" sama "trained_IVF1958_Flat_nprobe_1_Itsuki_v2.index"

@ryusuikagaku

8 ай бұрын

@@Ritsu_subhi corrupt mungkin? Kalau model lain indexnya bisa ga?

@ryusuikagaku

8 ай бұрын

@@Ritsu_subhi atau coba hapus yg trained dari zipnya

@Ritsu_subhi

5 ай бұрын

​@@ryusuikagakubang, ini yang bagian run gradio gak keluar publik link, emang lagi error kah?

Benerin dong bang code nya 😭 Error CUDA ini gabisa convert 😭😭😭

Might consider reporting the same problem to discord soon

no longer working..

kalau lagunya duet gimana kak?

@ryusuikagaku

8 ай бұрын

inference 2x pakai 2 model yang berbeda. download folder hasil inference di folder song_output. edit manual bagian duetnya di audio editor kesukaan kalian

hallo bang. saya masih belajar karna mulai tertarik. cuman kenapa setiap saya upload lagu dan ingin di convert dengan model ai yang sudah saya buat, selalu error bahkan berakhir gagal. mohon solusinya. terima kasih.. btw, kontennya sangat mudah dipahami dan menarik.. :)

@ryusuikagaku

7 ай бұрын

Codenya lagi error karena google colab ngeupgrade versinya dan ga compatible sama bbrp library. Nanti aku perbaiki

@ONEULPOP

7 ай бұрын

@@ryusuikagaku izin nambahin bang: issuenya muncul cuma pas misahin vocal dari lagunya. kalo inferencing pake stems yg udah ada sebelumnya (opsi 'install to drive' saya centang btw) gaada masalah, langsung muncul hasilnya

@ryusuikagaku

7 ай бұрын

Itu mungkin karena kamu pake lagu yg udah pernah diinference, kalau lagu baru pasti error juga wkwk. Katanya sih colab memperbarui versi CUDAnya, jadi error sama bbrp library

@thanh2425

7 ай бұрын

​@@ryusuikagakuAny way to actually make it usable while the code is being fixed?

@ryusuikagaku

7 ай бұрын

nothing can we do, only 2 option. downgrade cuda on colab or wait until onnxruntime support cuda 12

kak code nya apa?

kalau pake cara ini pake model udah jadi/model milik orang lain (bukan kita yang bikin) bisa gak bang?

@ryusuikagaku

8 ай бұрын

Bisa, tinggal di download modelnya

@Justfanmade

7 ай бұрын

@@ryusuikagaku tapi bang, kadang kadang kalau ai cover ada beberapa part yang ga ter cover, masuk nya ke backup vocal, gmn tuh bg?

@ryusuikagaku

7 ай бұрын

@@Justfanmade keberuntungan aja sih itu 😅

Bang pas generate stuck di 10% ada solusinya kah?

@thanh2425

7 ай бұрын

Did it also come with the "Failed to create CudaExecutionProvider" as well?

@Taeviel_Zhan

7 ай бұрын

@@thanh2425 Nope

@thanh2425

7 ай бұрын

@@Taeviel_Zhan The warning doesn't show up on the gradio link. It only pops up in the colab link. Everyone to has the same problem right now

@Taeviel_Zhan

7 ай бұрын

@@thanh2425 Oo i didn't check on colab

5:52 udah tak masukin link gdrive malah error bang

@ryusuikagaku

6 ай бұрын

- pastiin linknya ga private - pastiin itu link zip bukan link folder - pastiin isi dari zipnya

@chavezzx

6 ай бұрын

iya kak bener saya juga ga privat maupun dari download ataupun training sendiri tetep gabisa, tulisanya error no such file padahal udah lengkap dan public juga, kirain cuma saya yg kena problem kaya begini 😢

@Brotgaming2

6 ай бұрын

@@chavezzx lewat link huggingface juga gitu. Tulisan errornya "file format is not zip" padahal udah format zip

@chavezzx

6 ай бұрын

​@@Brotgaming2tunggu aja lah nanti juga diperbaiki

Bang, tarikan penyanyi AI Model nya gabisa ditiru ya bg? Ane ngetrain suara Ariel peterpan yg kebanyakan suara Dada, pas di ai cover ke lagu lain kok suara Perut semua 😭

@ryusuikagaku

7 ай бұрын

Aku ga paham suara dada suara perut wkwk

@muhammadraihan8832

7 ай бұрын

@@ryusuikagaku maksudnya teknik suaranya bang

@muhammadraihan8832

7 ай бұрын

@@ryusuikagaku pas cover penyanyi cewek, ane turunin pitch nya ke -12, pas bagian reff nya, kok suara ai nya ga mau naik ya? Ga bisa nada tinggi maksudnya, padahal di dataset lengkap ada suara tinggi dan rendah juga 🤔

@ryusuikagaku

7 ай бұрын

Mungkin penyanyi aslinya ga nyanyi di pitch yg sama, tapi berubah" di tiap part

Bang,kalo muncul no cuda gpu are available,itu kenapa y?

@ryusuikagaku

7 ай бұрын

Di colab bagian atas, ke runtime -> runtimr type -> nyalain GPU T4. Kalau gabisa berarti akunmu udah kena limit, tggu 24 jam atau ganti gmail

@halo1366

6 ай бұрын

​​@@ryusuikagakukalo akun yg dh kena limit, masih bisa dipake lagi kedepannya atau mau nunggu?

@ryusuikagaku

6 ай бұрын

Tggu 24 jam

bang kalo error that song inpu field . . gimana tu bang?

@ryusuikagaku

8 ай бұрын

Link lagunya dari youtube ga? Lagu lain bisa ga?

@andikurniawan4456

8 ай бұрын

dari youtube, surat cinta untum starla bang

@ryusuikagaku

8 ай бұрын

Cek di colab bagian muncul public url, disana harusnya ada detail errornya (scroll kebawah)

Ini bisa di hp ga bang?

bng udh gk bisa lagi kemaren kemaren bisa hari ini gw cobain lagi udh gk bisa jir tadi gw cobain lagi tiba tiba bisa 🪨

@ryusuikagaku

8 ай бұрын

lah aku baru make 30 menit lalu bisa

@rimuruxd6959

8 ай бұрын

@@ryusuikagaku pas gw reset laptop gw trus masuk linknya ada tulisan ini bng "Unable to render code block" Eh malah udh bisa lagi 🗿 "Edit"

@rimuruxd6959

8 ай бұрын

@@ryusuikagaku jaringan doang x ya yang lag

pitch nya ga bisa jadi 0 ya ? cuma 1 / 12 aja

@ryusuikagaku

8 ай бұрын

1/12 itu maksimalnya, tetep bisa 0

@ayanx234

8 ай бұрын

@@ryusuikagaku biar bisa 0 digimanain ya? Pilihannya cuma ada 1 / 12

Kok di gua muncul filed to load library bang🙏🙏

@ryusuikagaku

5 ай бұрын

kirim ss error ke discord

@Feldway5

5 ай бұрын

Udah bisa bang terima kasih , tutorialnya sangat bermanfaat 👍

"No CUDA GPUs are available", ini kenapa ya??

@ryusuikagaku

6 ай бұрын

Kena limit mungkin akunmu, hari ini udah ada pake lebih dari 3 jam? Tggu 24 jam atau ganti gmail lain

@muhammadfajarfirdaus2751

6 ай бұрын

@@ryusuikagaku walah, ini aku berlangganan colab pro ngaruh gk bro? apakah tetap 3 jam aja pemakaiannya?

@ryusuikagaku

6 ай бұрын

Ngaruh

Ko link nya ga muncul bang pas di run all

@ryusuikagaku

6 ай бұрын

Udah ditunggu?

@dandipermadi1230

6 ай бұрын

@@ryusuikagaku udah ttp aja link nya ga muncul

@dandipermadi1230

6 ай бұрын

@@ryusuikagaku di Run webUi nya ga muncul muncul dah 1 jam malah

@ryusuikagaku

6 ай бұрын

Ga selama itu juga sih, 10menit biasanya

Banh ko punyaku local ama public link nya nga muncul?nge stuck di 'can't import name' 'set_documentation_group'

@ramadhan2335

5 ай бұрын

Sama, punyaku muncul kayak gitu

@chriscanute

5 ай бұрын

Sama cuyy

@ramadhan2335

5 ай бұрын

Udah jumpa solusi Selepas dapat error kayak gitu buat command ini di collab: pip install gradio_client==0.8.1 Nanti udah bisa buka link

@Irish_Primrose

5 ай бұрын

Di codingan mananya itu? Edit:kaga usah,udh kuperbaiki,trims buat solusinya

@albedo_alchemy01

5 ай бұрын

​@@ramadhan2335 di bagian +Kode bukan bang buat Codingnya

jadi total nya ada berapa cara inference bang ? :V

@ryusuikagaku

8 ай бұрын

banyak, tiap orang pada bikin modifikasi mereka

tutorial nyak gak manfaat,,pas download model di link nya eh malah eror bacaan nya "tidak ada file seperti itu" aneh tutorial gak manfaat sama sekali

Ini collab nya harus pro ga bang?

@ryusuikagaku

8 ай бұрын

Ga

Bang kalau mau cover lagu duet yang suaranya lagi barengan gimana caranya

@ryusuikagaku

8 ай бұрын

inferennce 2x pake model yang berbeda. download hasil inferennya yang tanpa instrument. gabungin secara manual pakai audio editor yang kamu punya

Proses downgrade nya emang lama y bang?

@ryusuikagaku

6 ай бұрын

Lumayan

@Brotgaming2

6 ай бұрын

@@ryusuikagaku pas donlot model kok error y bang