Higgs Audio V2 Tokenizer
This is the tokenizer for the Higgs Audio V2 text-to-speech model.
Description
The Higgs Audio V2 Tokenizer handles text preprocessing and tokenization for the Higgs Audio V2 TTS model.
Model Details
- Model Type: Text Tokenizer
- Framework: HuggingFace Transformers
- Purpose: Text preprocessing for TTS
Usage
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("Toowired/higgs-v2-tokenizer")
# Tokenize text for TTS
text = "Hello, this is a test."
tokens = tokenizer(text, return_tensors="pt")
Features
- Optimized for English text
- Handles special characters and punctuation
- Compatible with Higgs Audio V2 LLM
- Fast tokenization for real-time applications
Intended Use
- Text preprocessing for text-to-speech
- Tokenization for Higgs Audio V2 model
- Research in TTS preprocessing
Technical Details
- Vocabulary size optimized for speech synthesis
- Special tokens for audio markers
- Efficient encoding/decoding
License
Please check the original model repository for license information.
Contact
For questions about this tokenizer, please contact the repository owner.