T

Tts En Fastpitch

Developed by nvidia
FastPitch is a fully parallel Transformer-based text-to-speech model capable of controlling pitch and phoneme duration, generating high-quality American English speech.
Downloads 4,701
Release Time : 6/28/2022

Model Overview

A Transformer-based parallel TTS model that generates expressive speech by predicting pitch contours, supporting real-time speech synthesis.

Model Features

Fully Parallel Architecture
Transformer-based fully parallel design for efficient speech synthesis
Pitch Control
Predictable and adjustable pitch contours for more expressive speech
Real-time Synthesis
Higher real-time factor compared to traditional Tacotron2 models
Unsupervised Alignment
Uses unsupervised speech-text aligner to improve synthesis accuracy

Model Capabilities

English Text-to-Speech
Pitch Control
Real-time Speech Synthesis
Mel-spectrogram Generation

Use Cases

Speech Synthesis
Voice Assistants
Generate natural and fluent speech responses for virtual assistants
Produces expressive American English speech
Audiobooks
Convert text content into speech for audiobook production
Adjustable pitch and speech rate for enhanced listening experience
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase