README.md

---
license: apache-2.0
datasets:
- Bingsu/KSS_Dataset
language:
- ko
base_model:
- microsoft/speecht5_tts
pipeline_tag: text-to-audio
---

# Korean SpeechT5 (Jamo Tokenizer, KSS)
If you use this model in research or production, or further fine-tuning,
please cite:

@misc{ahnhs2k_speecht5_korean,  
  author = {Ahn, Hosung},  
  title = {Korean SpeechT5 TTS Model},  
  year = {2025},  
  publisher = {Hugging Face},  
  url = {https://huggingface.co/ahnhs2k/...}  
}

## 모델 특징
- Base Model: microsoft/speecht5_tts
- Dataset: Bingsu/KSS_Dataset
- Tokenizer: Jamo-based Korean tokenizer (character-level)
- Speaker Embedding: microsoft/wavlm-base-plus-sv
- Vocoder: microsoft/speecht5_hifigan
- Sample Rate: 16 kHz
- 단일 화자 한국어 TTS 모델