mah92/hifigan_for_sherpa
By mah92 · ↓ 0 · ♥ 0 · Updated 2025-05-08
What's in this repo
This page lists the files in mah92/hifigan_for_sherpa, including hifigan_universal_22050hz_v1.onnx, hifigan_vctk_22050hz_v1.onnx, hifigan_vctk_22050hz_v2.onnx and more.
- 4 ONNX models (.onnx files)
- 10 files, including 4 .onnx, 4 .wav, 1 .md
- Last updated 2025-05-08
ONNX files
| File | Size | Download |
|---|---|---|
| hifigan_universal_22050hz_v1.onnx | — | Download |
| hifigan_vctk_22050hz_v1.onnx | — | Download |
| hifigan_vctk_22050hz_v2.onnx | — | Download |
| hifigan_vctk_22050hz_v3.onnx | — | Download |
How to use
These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.
README
بسم الله
ATTENTION: HIFIGAN IS RETIRED. USE VOCOS INSTEAD.
-
Converted universal v1, v2, v3 22050Hz hifigan models from here and here
-
Converted to onnx using this
-
Added metadata as such using the above python script: {'model_type': 'hifigan', 'model_filename': 'hifigan_universal_22050hz_v1', 'sample_rate': 22050, 'version': 1, 'model_author': 'jik876', 'maintainer': 'k2-fsa', 'dataset': 'Universal', 'url1': 'https://github.com/jik876/hifi-gan', 'url2': ''}
Note1: hifigan v2 is much smaller, faster and with lower memory need
Note2: As LJ speech is a female single speaker dataset, hifigan based on it does not work well on men's voice. vctk models work well for both genders.
Conclusion: Use vctk_v2!
Model info
Created: 2025-02-11
Updated: 2025-05-08
View on Hugging FaceFiles (10)
Related models
More from mah92
More models
No task type declared.