lmnt-com/diffwave
A diffusion-based neural vocoder that converts Gaussian noise into high-quality speech waveforms conditioned on Mel spectrograms.
Not currently ranked — collecting fresh signals.
star history
DiffWave is a fast, high-quality neural vocoder and waveform synthesizer built with PyTorch. It uses a diffusion probabilistic model that iteratively refines Gaussian noise into speech waveforms. The model can be conditioned on log-scaled Mel spectrograms for text-to-speech synthesis or run unconditionally for raw waveform generation. It supports fast sampling, mixed-precision training, and multi-GPU training, with pretrained models available for immediate use.
Frequently asked
- What is lmnt-com/diffwave?
- A diffusion-based neural vocoder that converts Gaussian noise into high-quality speech waveforms conditioned on Mel spectrograms.
- Is diffwave open source?
- Yes — lmnt-com/diffwave is open source, released under the Apache-2.0 license.
- What language is diffwave written in?
- lmnt-com/diffwave is primarily written in Python.
- How popular is diffwave?
- lmnt-com/diffwave has 885 stars on GitHub.
- Where can I find diffwave?
- lmnt-com/diffwave is on GitHub at https://github.com/lmnt-com/diffwave.