Light

sadeghkrmi / pertts-streamlit Goto Github PK

View Code? Open in Web Editor NEW

13.0 2.0 1.0 111.5 MB

Persian text-to-speech streamlit interface

Home Page: https://tts.datacula.com/

License: Apache License 2.0

Python 75.00% CSS 3.56% Dockerfile 21.44%

machine-learning persian-dataset persian-language persian-tts text-to-speech tts-model vits deep-learning

pertts-streamlit's Introduction

pertts (persian text-to-speech)

This is the implementation and a web interface for pertts (Persian text-to-speech)

powered by piper

the live version of persian tts called pertts

with love from datacula.com

Voice: We are using an AI-based TTS system, trained with Amir Sooakhsh's voice from rokhpodcast, Special thanks to Amir :)

dataset

https://huggingface.co/datasets/SadeghK/datacula-pertts-amir

🛠️ Installation

docker

Build with docker from scratch and run

docker build --no-cache -t pertts:1.0 .
docker container run --name st --rm -it -p 8501:8501 pertts:1.0

Run the latest version of the docker image from docker hub

docker image pull sadeghk/pertts
docker container run --name st --rm -it -p 8501:8501 sadeghk/pertts

python

install piper-tts using pip and download the model in pertts-streamlit/model directory

`` pip install piper-tts

and then run
```bash
echo 'سلام و درود بر همه فارسی زبانان' | piper \
  --model epoch=5261-step=2455712.onnx \
  --output_file dorood.wav

Windows

download the executables for windows piper_windows_amd64.zip from piper, and unzip go to the piper directory where piper.exe exists and create a folder with the name models. Download the model for Persian/Farsi from huggingface with name fa_IR-amir-medium.onnx and fa_IR-amir-medium.onnx.json to models directory.

open a PowerShell and cd to the directory where piper.exe exists, and run

echo "سلام و درود بر شما" | .\piper.exe --model .\models\fa_IR-amir-medium.onnx --output_dir .\outputs

pertts-streamlit's People

Stargazers

Watchers

Forkers

pertts-streamlit's Issues

Dataset creation

@SadeghKrmi , Hello

Could you please a share a tutorial how did you created your dataset ?

This was amazing

Sorry to bother you but I just wanted to share my excitement 😆
For the first time since 15+ years ago, we finally have access to a very high quality, open-source Persian text-to-speech. This is a very nice model and I'm blown away by the model's small size considering its accuracy.
Thanks for maintaining this project, I love the Web UI! 👍🏻

Help for creating dataset

Thank you @SadeghKrmi , i wanted to know these about your dataset creation process :

how did you get transcription of rokhpodcast.ir audio tracks?
how did aligned voice with transcription?
Size of dataset?
tools you used for creating your dataset?

helppppp!

my problem is when i want use the powershell to send farsi phrase to piper, the powershell does not support the persian unicode and i want to know that you proceed this project, how you overcome this because i cannot find any solution i the internet.

speech sometimes differs from the actual text

Great job in creating the first applicable open-source AI text to speech for Persian language which has excellent voice quality.
I just relized that the amir model in this initial release sometimes sometimes ignore the actual text and read it as a conversational text.
This is not what we expect from a text to speech. I belive this problem can be solved by adjusting the text based on what actually was read by the voice talent.
This issue is especially problematic if a blind user of computer want to use this text to speech during editing of a document. Imageine what happen!

Input text:
رنگین‌کمان پدیده‌ای نوری و کمانی است که زمانی که خورشید به قطرات نم و رطوبت جو زمین می‌تابد باعث ایجاد طیفی از نور در آسمان می‌شود.

Audio text:
رنگین‌کمان پدیده‌ای نوری و کمانیه که زمانی که خورشید به قُطرات نم و رطوبت جو زمین می‌تابد باعث ایجاد طیفی از نور در آسمان می‌شه.

I could contribute toward solving this issue but don't know where to find the training dataset!

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.