Downloads · 30 days
0
jjiang4/thorsten_hifigan_ft_ljspeech
thorsten_hifigan_ft_ljspeech is a text-to-speech model from jjiang4. Use it when you need text read aloud. It is set up for espnet. The card lists the license as cc-by-4.0.
HiFi-GAN V1 vocoder for the German Thorsten-Voice Dataset 2022.10 speaker, fine-tuned for 40k steps from parallelwavegan/ljspeechhifigan.v1 with kan-bayashi/ParallelWaveGAN.
Downloads · 30 days
0
Access
Public
Updated Aug 27, 2026
Repo size
55.8 MB
Likes
0
Public
Click a slice to open those files.
.pkl55.8 MB · 100%
From the Hugging Face model README
jjiang4/thorsten_hifigan_ft_ljspeechHiFi-GAN V1 vocoder for the German Thorsten-Voice Dataset
2022.10 speaker, fine-tuned for 40k steps
from parallel_wavegan/ljspeech_hifigan.v1 with
kan-bayashi/ParallelWaveGAN.
Intended for the ESPnet egs2/thorsten/tts1 recipe. It replaces Griffin-Lim and
raises UTMOS from 1.42 to 3.04 on that recipe's test set.
Matches egs2/thorsten/tts1's feats_extract_conf exactly, so it consumes the
recipe's mel-spectrograms unchanged:
| sampling rate | 22050 |
| FFT size | 1024 |
| hop size | 256 |
| mel bins | 80 |
| fmin / fmax | 80 / 7600 |
checkpoint-40000steps.pkl — generator weights. Training state (discriminator
and optimizers) is stripped; the output is bit-identical to the full
checkpoint.config.yml, stats.h5 — read by ESPnet when loading the vocoder. Keep all
three files in the same directory.cd espnet/egs2/thorsten/tts1
./run.sh --stage 8 --inference_model valid.loss.best.pth \
--vocoder_file /path/to/checkpoint-40000steps.pkl
from espnet2.bin.tts_inference import Text2Speech
tts = Text2Speech.from_pretrained(
"jjiang4/thorsten_tts_train_tacotron2_raw_phn_espeak_ng_german",
vocoder_file="/path/to/checkpoint-40000steps.pkl",
)