AIbase
EN
Home
Tts Ru Hifigan Ruslan
T

Tts Ru Hifigan Ruslan

Developed by bene-ges
A Russian text-to-speech model trained on the RUSLAN corpus, using FastPitch and HifiGAN architectures, supporting speech synthesis at 22.05kHz sampling rate.
Speech Synthesis Other#Russian speech synthesis#22.05kHz high sampling rate#Single male speaker
Downloads 38
Release Time : 4/18/2023

Model Overview

This model is a Russian text-to-speech (TTS) system capable of converting Russian text into natural speech. It uses IPA phonetics for text preprocessing (G2P), generates mel-spectrograms via FastPitch, and synthesizes high-quality speech with the HifiGAN vocoder.

Model Features

High-quality speech synthesis
Uses the HifiGAN vocoder to generate high-quality speech at 22.05kHz sampling rate.
Phonetic preprocessing
Employs IPA phonetics for text preprocessing (G2P), improving pronunciation accuracy.
Single-speaker model
Trained on the RUSLAN corpus, focusing on single male-voice speech synthesis.

Model Capabilities

Russian text-to-speech
22.05kHz high-quality speech synthesis
IPA-based phonetic conversion

Use Cases

Speech synthesis applications
Audiobook generation
Converts Russian text into natural speech for audiobook production
High-quality speech output at 22.05kHz sampling rate
Voice assistants
Provides speech synthesis capabilities for Russian voice assistants
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase