B

Bigvgan 22khz 80band

Developed by nvidia
BigVGAN is a universal neural vocoder achieved through large-scale training, capable of providing high-quality audio output for tasks such as speech synthesis.
Downloads 2,344
Release Time : 7/15/2024

Model Overview

BigVGAN is a universal neural vocoder that achieves high-quality audio generation through large-scale training and is suitable for tasks such as speech synthesis.

Model Features

Large-scale training
Trained on a large-scale dataset to provide high-quality audio output.
CUDA kernel fusion
Implement fully fused CUDA kernels for anti-aliased activation to improve inference speed.
Multi-sampling rate support
Supports sampling rates up to 44 kHz and upsampling rates of 512x.
Improved discriminator
Trained using a multi-scale subband CQT discriminator and multi-scale mel spectrogram loss.

Model Capabilities

High-quality audio generation
Speech synthesis
Multi-sampling rate support

Use Cases

Speech synthesis
Text-to-speech
Convert text into natural speech
High-quality speech output
Audio enhancement
Improve the clarity of low-quality audio
Improved audio quality
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase