Відео

I tired this out on a rtx 3600 12gb model and it's fast. Quicker than speaking, maybe 2x faster to process than to listen to. Sounds really good to me.

can we run piper tts in gpu using cuda ?

This error originates from a subprocess, and is likely not a problem with pip. error: subprocess-exited-with-error im getting this error please help someone

Hello Thorsten, thanks for your great channel. I came about these videos which shows how one can train F5 with different languages ua-cam.com/video/UO4usaOojys/v-deo.html ua-cam.com/video/RQXHKO5F9hg/v-deo.html As you are experienced with training of speech models, I am wondering how much hours material would be required to train a German language model in good quality and what things should be considered in regards to training data. In the referenced youtube video the creator simply takes audiobooks. Can one expect to get a good quality model in this way?

Hi Thorsten, thank you for another excellent tutorial. I have installed f5 on a Raspberry Pi 5 and it generates very good quality output but to be expected it is very slow. I am trying to understand how f5 works, does it take a standard model and modify it in some way using the ref_text & audio before generating the desired output (gen_text)? Is there an intermediate stage that could be executed separately? Thanks Ernie

How much time it took to train one model??? And how powerfull is your system?

If anybody is encountering errors when installing TTS, try pip install coqui-tts (tts is deprecated as of november 2024)

I went down this rabbithole myself last year. Alas! Have I watched this 6min video I would've saved lotsa disk space and time (weeks). Cheers! ✌

Hi, Thorsten, the community thrives because of people like you - thanks for your work!

thanks! it is faesabel to do all of that trought scripted pyton code?

the greetin comes from hessen in germany right? :D funny intro and exactly what i am searching for :) abo

What GPU do you have on your computer?

May I ask what gpu you are using, or if it is using a gpu?

For anyone coming recently, the tts repo isn't maintained anymore according to an issue post on the github. It results in an error when running 'pip install tts'. This fork worked for me instead: 'pip install coqui-tts'

is there a way to use this with nvidia gpu on windows to speed up performance?

can this be deployed and hosted on a server?

I tried it and it works but it did not sound like me. Nothing close to what you did. Not a fan at this time it really should have done better. Thanks for sharing you got my thumbs up...

great stuff!

Haha the F5 joke😂. The progress is amazing, right? Still waiting for german support for F5... Anyway in english it is now already easy to create synthetic voice datasets for piper for example, just an idea😊

i need more georgian voices. there is only Natia(woman) voice ( how can i make them? any tut?

Thanks for your video. F5 TTS is absolutely stunning! Let's hope they will include other languages (GERMAN) soon. ;)

That was great!! Thanks for your content! I've got this running now and it is amazing!!

That whisper at the beginning really sounded like Stephan Molyneux?!!!

Great

Try MaskGCT

You made a reference to your computer speed. Care to elaborate on its GPU and CPU and ram?

I enjoyed the intro it made me laugh.

Hi Thorsten! I want to make a portfolio website where people can talk to myself. Id have a text to text that knows everything about me and that would go to a tts of my own voice to tell it what to say each time. My problem is hosting. I dont understand how the APIs of these tts models work and how id be able to host it as most gpu hosting websites offer per hour rates which seem very expensive.. what do i do! maybe ive got the wrong approach..

1.5x speed is about right 😂

ChatGPT was right? 🥴

Could you please help me to decide what model(s) (will use as voiceover) is fit, more natural human sound for faceless YT videos?

Spitzen Arbeit 👍 Es ist eigentlich genau das was ich gesucht habe. Bin jedoch absoluter Computer Legasteniker und hoffe das mir ein Freund dabei hilft deine Stimme zu installieren.

Thank you very much for this excellent and very easy to follow tutorial!

I can't get good results with the Parler TTS demo whenever I let it speak a paragraph. It starts out fine but then it starts to drop more and more words, garbling up all the text into an incomprehensible mess. Very weird.

Love thi channel 😊😊😊

Really helpful video, thanks!

Danke für's video. Es funkioniert endlich! Richtiger ehrenmann 😀

Sehr gutes Video. Hätte ich gewusst, dass du hier die Installation auf Windows vornimmst, hätte ich mir 2 Tage arbeit gespart :D

hi Thorsten, may the next "how to" would be training coqui-TTS model based on Glow-TTS and HiFiGAN vocoder?

Hi Thorsten, How many hours/steps you spent to trains your DE dataset to become usable model in couqi-tts? I'm trying to do some model training with my dataset (35 minutes of audio) and I start hearing some voice on 10k steps but it is far away from what I would like to get....

How can i be sure that all wavs are used in trainning process? For test i have 650 segments ( 35 minutes ) but i would like to prepare several hours but do not know if it is needed... have you have any sugession?

What is the different between vocoder and tts? Should i train vocoder and then model based on the same dataset wavs? Does the model with vocoder be better than without it?

Hi would it be possible to you to create traing vocoder HiFi-GAN?

Hello Thorsten, could you please create video in which you can shows different between: tacotron2-DDC, vits_tts, glow_tts....?

I am severly dissapointed as I just wanted to install my native tounge HEBREW to be used at Firefox read aloud add on. I scrolled my way up down than upwards the scrollrd down again with thay stronge negative feeling i get as a jewish woman born at my only known country Israrel, as my family elders may they rest in peace sufferered the absolut horors of ww2 holocaust somehow managed to survive what so many other could not. Did nt. And this notion of how so very real that was not back in the far past. but i cannnot feel that. Im a refuge of horror against each and every 3rd gen jew. I got that notion so vivid and real in me. The dear aunt who was my nanny when i was growing up for so many years My dear nanny rip aunty paula her hand holding mine as we cross the road as she tool me day by day to school. ... . That damn number i never ever tryed to memorized refused in my subcontius to place thos4 horrible numbers. But their damn geeanish turqoize bluish shade is carved in my mind always a testemony to the hell thay was in her life. I cant realy get how they managed to build a life. But that notion of hate and impulse of enilating me as a jew born this way..... is so much present. 7.10.2023 2nd holocaust came upon us. HAMMAS was and is since its creation driven to enielate a deatroy Israel, and was inflicting TERROR day in and out back at those days when Israel was stupied enough to credify the up until a moment b4 ad after Oslo to recognize the PLO as a form of national entiity. The fisrt nation that did this was Israrel. Israel forcefully painfully removed by force all the lives and livelihoof that was existing i damn Gaza strip. After removing our own brothers and sistets and childrrn that was living at modest areas growing the best herbs i can recall employing the local ppl of Gaza. Israel the country destroyed to rubbles whatever was made by Israel. Living it so very purged of the presensce of Israel. What better statement was it for the ppl that were ppl of Gaza back than. Here its all damn your. Our nation was torn by that move. But as elections took place in Gaza. GAZA chose Hammas. The ppl chose Terrorist leadership masking as the good faithfull religious supposed to be better leadership that will take care of the poor people that the pillow was taking its money for you know the exact same thing but way more violent way more terrorist that later on Hamas leadership to the same money and used it for making Gaza ‏into Hamastan old regime in Gaza became under the control of a terrorist organization or people if you may call those who preached to my death human ‏became a machine supporting a terrorist organization Iron Dome for Israel is no less than a miracle but every attack ‏made by them targeting rockets and missiles toward first southern cities..than enlarging bit by bit the distance. I was and still am a valid target. The ppl choose Tha ppl agree The ppll came upon our nation on that second holocaust happend at 7.10 But not only Gaza i speaking arabic which i honestly regreting not learning it ..... Well What about Iran? Sure. I support the ppl . I have a gut feeling that sometday we shall discollver the unbareable amount of our both cultures... Yet The Natzi regime states over and over indoctrinate a younger future genaration to destroy ISRAEL Nazis Propoganda education Millitary developm÷nt declare to destroy my home a 2 other jewish elder kingdoms who rulled this troublea filled land I complain No language available for my nation . No Hebrew. How come? Why am my tounge and nation existing just like other nations got their own I insist HEBREW Should be plantifull with at least 5 voices to choose from I have a languge that might be lacking constant usase but ots alive and kicking I am without being asked a human in a cointry called Israel . Not all Israelis are jews. And israel citizens are not living at aparthide. We live in something more like tribes and we to my view should mingle way more between outself. Less apart But all Isralis call and know that our imperfect collective home is our home. Please make Hebrew planty of speach engines ASAP please Please make it happen. We are no less that other nationalities. We try our best at a biased world

Hello Thorsten, have you had time to implemented this: "textCleaned": text.lower() # TODO: Add textcleaner library (multilanguage support) in your script?

I'm also interested in Voice Clone, please do share your knowledge in coming videos

Hello Thorsten, does your script include specify model name that whisper should use and language for the model? can you update it? How can i use it on my RTX card? update: i have updated your script to use RTX by adding this code: device = "cuda:0" if torch.cuda.is_available() else "cpu" torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32 model.to(device) but first torch with cuda enabled needs to be installed....

What I really need is the ability to create Unreal Engine projects that do what Convai can do, but run locally within the project, so I don't have to pay a lot of money. Ideally running Gemma 2 9B and a good TTS with the possibility for voice cloning that are all part of the Unreal Engine game project.

Hi! Thanks for the video, it was really helpful for anyone looking to get into voice cloning. However, I have a question. I followed all the steps, and I noticed that after exporting, the .wav files generated in the wavs folder have a sample rate of 48000 Hz. However, for fine-tuning a pre-trained Piper model, it seems that .wav files with a sample rate of 22050 Hz are required. My question is: are the 48000 Hz audio files still acceptable for fine-tuning, or should I convert them to 22050 Hz before proceeding?