IndexTTS2 SECourses Premium App Just Got Published 1 Click to Install and Voice Clone and Generate

IndexTTS2 SECourses Premium App Just Got Published - 1-Click to Install and Voice Clone and Generate

Full tutorial link > https://www.youtube.com/watch?v=YbgFVKWB7hs

Download and install app from here : https://www.patreon.com/posts/IndexTTS2-SECourses-Premium-Voice-Clone-App-139297407

This app is built upon IndexTTS2 open source model. And this is a 0-shot voice cloning. Instant voice cloning without any training or LoRA etc. Just used an average quality 22 second audio reference. You will see it at the end of the video.

Hopefully full tutorial coming soon

The app works locally on Windows, and Linux

Also you can run on Massed Compute and RunPod as usual if you are GPU poor

As low as 8 GB GPUs should run fairly well, lower may work too

Video Transcription

• 00:00:00 Today I am going to introduce  you Index TTS Premium. Built  

• 00:00:04 by SECourses. 1-click to install on  Windows, RunPod and Massed Compute.

• 00:00:10 We handle all the installation  for you with all pre-compiled  

• 00:00:12 libraries so it works amazing on  all GPUs such as RTX 2000 series,  

• 00:00:17 3000 series, 4000 series or 5000 series.  We support literally every GPU out there.

• 00:00:24 With as low as 8 GB GPU, you can run  this amazing app on your local computer  

• 00:00:30 and generate amazing extremely  high quality cloned voice audio.

• 00:00:34 Cloning voice with this model  is insanely easy and also good.

• 00:00:38 This is the version 1 release of the app and  hopefully many more features coming soon.

• 00:00:44 For example some of the upcoming  features are as: save and load presets,  

• 00:00:49 number of multiple generations, saving  full metadata of each generation,  

• 00:00:53 continue generating a big text, give  reference audio from path and such.

• 00:00:57 Since this app is so good at the moment,  

• 00:01:00 this will be our number 1 text to speech  app for a while until a better one arrives.

• 00:01:04 With this amazing app, you  can literally generate an  

• 00:01:07 entire audio book with a single click of a button.

• 00:01:10 It automatically intelligently splits given  text into parts and process every one of them.

• 00:01:16 When you provide a very long text, it will process  it part by part and combine audio at the end.

• 00:01:22 The audio and voice will be extremely coherent  

• 00:01:24 and stable. That is what makes this  app and model different than others.

• 00:01:29 Also if you want a very clean voice,  

• 00:01:31 give as an input a very clean voice.  It makes huge difference and impact.

• 00:01:36 The length of the text doesnt matter  since it is processed as chunks.

• 00:01:39 All generated voices will be saved  inside outputs folder automatically.

• 00:01:43 For lower memory GPUs, enable  Low Memory Mode checkbox.

• 00:01:48 The speed of the generation is also insane. It  is like 100 seconds audio generation takes 25  

• 00:01:54 seconds. Of course this depends on GPU but  you can get this speed with consumer GPUs.

• 00:02:01 Now i will play the original voice  used for voice cloning to generate  

• 00:02:06 this audio so you can see how  high quality this speech was.

• 00:02:09 I have also provided this used reference audio as  

• 00:02:12 a demo audio in the zip file  so you can use and replicate.

• 00:02:15 Original audio is a real audio taken from  an interview. the uh Neuralink device is  

• 00:02:20 kind of like a a fitbit or an Apple Watch um  that's uh where where we we we take out a a  

• 00:02:28 sort of a a small section of skull about the  size of a quarter, um replace that with uh  

• 00:02:32 what in many ways really is very much like um uh  Fitbit, Apple Watch or or something like that.