qwen3_tts_tokenizer_12hz¶


Model type	`qwen3_tts_tokenizer_12hz`
Class	`Qwen3TTSTokenizerV2Model`
Task	`text-generation`
Source	`models/qwen3_tts_tokenizer.py`

Description¶

Composite Qwen3-TTS codec tokenizer model.

Holds both decoder and encoder sub-models. The composite preprocess_weights routes HF weights and precomputes VQ codebook embeddings from embedding_sum / cluster_usage.

HF class: Qwen3TTSTokenizerV2Model in qwen_tts.

Parameters: config: Architecture config.

Usage¶

mobius build --model <MODEL_ID> output_dir/

from mobius import build

model = build("<MODEL_ID>")