F5-TTS and E2-TTS - AI Model That Fakes Fluent Speech - Install Locally

[@fahdmirza]

🔥Install F5-TTS on Windows Locally - kzbin.info/www/bejne/ZnmTZp14m7aeodUsi=oSqBtL-2MR_zBdOv 🔥Train F5-TTS on Your own voice data - kzbin.info/www/bejne/iIK7eX6Faqtsnsksi=9U4AEyrDKI8IwXkK 🔥F5-TTS and E2-TTS - AI Model That Fakes Fluent Speech - Install Locally - kzbin.info/www/bejne/bGqcoamKjLOAidEsi=d7QIITlqxgHMNI7C

[@AdamNewcombMumbleTrash]

My dude, you have got to start selling "Let me clear the screen" T-shirts. YOU MUST!! I appreciate your work. Very informative.

[@fahdmirza]

Thank you

[@badrinarayanans355]

Tried it out, it's giving pretty good results. Looking forward to learn about finetuning this TTS models. Expecting a video from you.

[@fahdmirza]

Great to hear!

[@siddhubhai2508]

Sir there is a problem, now in the f5 repo there is no test inference python file, now what to do? Please make a new video no the updated version.

[@Homosapien77]

Depends if you give it isolated audio of a person talking like forty minutes worth of it the results are solid. If you are getting bad results just try getting longer footage removing backround noise and cleanup before inputting the mp3.

[@fahdmirza]

Sure, thanks.

[@john_blues]

They use a lot of non technical words to describe this(Fairytaler that Fakes Fluent and Faithful Speech with Flow Matching) but it seems like it's just a voice clone app. It's good. But, just voice cloning.

[@oskar4239]

Show me you have only cursory experience in TTS models without telling me you have only cursory experience in TTS models

[@john_blues]

@@oskar4239 And what's wrong with that? I'm here to learn.

[@fahdmirza]

Thanks for the feedback.

[@slamdunk0113]

Thanks a lot ~~ Does this F5-TTS have tutorials for generating other languages like Japanese?

[@fahdmirza]

Not yet, but watch this one for trying it out 🔥Install F5-TTS on Windows Locally - kzbin.info/www/bejne/ZnmTZp14m7aeodUsi=oSqBtL-2MR_zBdOv

[@ZergRadio]

Is it only for English voices or any language?

[@fahdmirza]

Chinese and English

[@ZergRadio]

@@fahdmirza wrote "Chinese and English", Thankz mate

[@TheGodParticles]

once you have done one process can this "model" or "dict" be used with open-webui for tts?

[@SiD-hq2fo]

will there be guide we expecting for finetuning this model ?

[@fahdmirza]

Noted.

[@richardrispoli4508]

Pretty good result.

[@fahdmirza]

I think so too!

[@vishalkhombare]

Thanks a lot!!

[@fahdmirza]

You're welcome!

[@theh1ve]

The generation times are quite slow, although i guess that is use case specific! I assume it can run on GPU and crankmout faster generations?

[@fahdmirza]

yes thats correct

[@NickMak-m2c]

The F5/E2 is too choppy and oddly broken up, it sounds like another language trying to be English. Thank you for the demo, mate!

[@fahdmirza]

Thanks for watching!

[@siddarth26]

Bro. Did it work without gpu

[@fahdmirza]

yes

[@sidarth404]

hey bro. can you run this in google colab.

[@fahdmirza]

No, but you can run it on Windows easily : 🔥Install F5-TTS on Windows Locally - kzbin.info/www/bejne/ZnmTZp14m7aeodUsi=oSqBtL-2MR_zBdOv

[@siddhubhai2508]

Do you know hindi?

[@fahdmirza]

No.

[@LeeBrenton]

seems like it's on pinokio

[@fahdmirza]

yes and I already did the video on it.

[@mahaltech]

hello i try to flow you on video , first repo is change its not the same even i recognize i have to run speech_edit.py the output is intersection its not clear voice its give half from original voice and contain small piece of my text that i want to generate to voice

[@Yuuyu_play]

😇😇😇😇

[@fahdmirza]

Thanks for the support.

[@mal-avcisi9783]

hey bro, can you give me 99% discount coupon for H100 ?

[@fahdmirza]

sure, why not

[@mal-avcisi9783]

@@fahdmirza thanks man very kind

[@siddhubhai2508]

Do you know hindi?

[@fahdmirza]

No.