HiFi-GAN is a generative adversarial network for speech synthesis. HiFi-GAN consists of one generator and two discriminators: multi-scale and multi-period discriminators. The generator and discriminators are trained adversarially, along with two additional losses for improving training stability and model performance.
The generator is a fully convolutional neural network. It uses a mel-spectrogram as input and upsamples it through transposed convolutions until the length of the output sequence matches the temporal resolution of raw waveforms. Every transposed convolution is followed by a multi-receptive field fusion (MRF) module.
For the discriminator, a multi-period discriminator (MPD) is used consisting of several sub-discriminators each handling a portion of periodic signals of input audio. Additionally, to capture consecutive patterns and long-term dependencies, the multi-scale discriminator (MSD) proposed in MelGAN is used, which consecutively evaluates audio samples at different levels.
Source: HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech SynthesisPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Speech Synthesis | 15 | 25.00% |
Text-To-Speech Synthesis | 6 | 10.00% |
Voice Conversion | 5 | 8.33% |
Voice Cloning | 2 | 3.33% |
Speech Enhancement | 2 | 3.33% |
Speaker Verification | 2 | 3.33% |
Dialogue Generation | 1 | 1.67% |
Domain Adaptation | 1 | 1.67% |
Self-Supervised Learning | 1 | 1.67% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |