lmnt-com/wavegrad

A fast, high-quality neural vocoder.

47
/ 100
Emerging

This tool helps turn audio descriptions, like a log-scaled Mel spectrogram, into realistic speech audio files. It takes a compressed representation of sound and reconstructs the full, high-fidelity waveform. Voice AI researchers, audio engineers, or developers building text-to-speech systems would use this to generate natural-sounding speech from spectral data.

296 stars. No commits in the last 6 months.

Use this if you need to convert Mel spectrograms into high-quality, natural-sounding audio waveforms for speech generation.

Not ideal if you're looking for a simple text-to-speech solution without needing to work with spectrogram data directly or if your primary need is general audio processing beyond speech.

speech-synthesis voice-ai audio-generation signal-processing text-to-speech
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 21 / 25

How are scores calculated?

Stars

296

Forks

51

Language

Python

License

Apache-2.0

Last pushed

Jul 18, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/lmnt-com/wavegrad"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.