Views
No views yet
magma90909/vocence_miner_v3
with the ckpt-800 LoRA adapter folded into the Talker weights.
No PEFT layers at runtime — load directly with Qwen3TTSModel.from_pretrained.1from qwen_tts import Qwen3TTSModel
2
3wrap = Qwen3TTSModel.from_pretrained("ratrys/sft-tts-800")
4wavs, sr = wrap.generate_voice_design(
5 text="Hello, this is a test.",
6 instruct="A young adult female speaker speaks calmly at a normal pace.",
7 language="english",
8 temperature=0.9, top_p=1.0, top_k=50,
9 repetition_penalty=1.05, max_new_tokens=600,
10)config.json, model.safetensors — Qwen3-TTS Talker + Code Predictor (merged)speech_tokenizer/ — 12.5 fps × 16 codebook neural codec (unchanged)tokenizer.*, vocab.json, merges.txt, added_tokens.json,
special_tokens_map.json — Qwen2 BPE tokenizergeneration_config.json, preprocessor_config.jsonvocence_config.yaml, chute_config.yml — runtime + Chutes deploy hintsminer.py into this repo (same contract as macminix's:
class Miner with __init__(path_hf_repo: Path), warmup(),
generate_wav(instruction, text) → (np.ndarray, int)); the standard
Vocence chute wrapper will load this model unchanged.merge_info.json for the exact base path, adapter
path, LoRA hyperparameters, and merge timestamp.