How to use amupd/hifigan with Transformers:
# Load model directly from transformers import SpeechT5HifiGan model = SpeechT5HifiGan.from_pretrained("amupd/hifigan", dtype="auto")
