Kikiri German — Victoria
German single-speaker TTS model fine-tuned on the Victoria Asztaller voice using StyleTTS2 Stage 2.
Built on top of kikiri-german-base-51speakers-synthetic.
Model Details
| Property | Value |
|---|---|
| Architecture | StyleTTS2 Stage 2 (Kokoro-compatible) |
| Language | German (de) |
| Speaker | Victoria Asztaller (synthetic) |
| Training samples | 455 |
| Training epochs | 10 |
| WavLM discriminator | active from epoch 4 |
| Mixed precision | bf16 |
| Sample rate | 24 kHz |
Demo
StyleTTS 2 — Abstract (German)
Usage
# Uses the kokoro library as underlying framework
from kokoro import KPipeline
pipeline = KPipeline(lang_code="de", model_path="kikiri_german_victoria_ep10.pth")
voicepack = pipeline.load_voice("voices/victoria.pt")
text = "Guten Tag, wie geht es Ihnen?"
audio = pipeline(text, voice=voicepack)
Files
| File | Description |
|---|---|
kikiri_german_victoria_ep10.pth |
Kokoro-compatible model weights (Stage 2, epoch 10) |
voices/victoria.pt |
Victoria speaker voicepack (512-dim style embedding) |
audio/styletts2_demo.wav |
Demo: StyleTTS 2 abstract read by Victoria |
License
Apache 2.0