Error loading model: ModelConfig.__init__() missing 11 required positional arguments
Hi guys.
Running the model I got back this error. Not cler what caused it. Any hints.
I am using M1 studio and python 3.12. Same error for all the MLX QWEN3-TTS models
python -m mlx_audio.tts.generate --model mlx-community/Qwen3-TTS-12Hz-0.6B-CustomVoice-bf16 --text "Hello, this is a test."
config.json: 2.34kB [00:00, 21.2MB/s] | 0/12 [00:00<?, ?it/s]
preprocessor_config.json: 100%|████████████████████████████████████████████████████████████████████████████| 127/127 [00:00<00:00, 2.94MB/s]
configuration.json: 100%|████████████████████████████████████████████████████████████████████████████████| 76.0/76.0 [00:00<00:00, 2.18MB/s]
generation_config.json: 100%|██████████████████████████████████████████████████████████████████████████████| 245/245 [00:00<00:00, 7.24MB/s]
model.safetensors.index.json: 32.3kB [00:00, 147MB/s] | 0.00/76.0 [00:00<?, ?B/s]
config.json: 5.85kB [00:00, 44.1MB/s] | 0.00/245 [00:00<?, ?B/s]
merges.txt: 1.67MB [00:00, 12.5MB/s] | 1/12 [00:00<00:10, 1.07it/s]
preprocessor_config.json: 100%|█████████████████████████████████████████████████████████████████████████████| 234/234 [00:00<00:00, 910kB/s]
tokenizer_config.json: 7.34kB [00:00, 4.48MB/s]
vocab.json: 2.78MB [00:00, 11.3MB/s] 0%| | 0.00/682M [00:00<?, ?B/s]
speech_tokenizer/model.safetensors: 100%|████████████████████████████████████████████████████████████████| 682M/682M [00:36<00:00, 18.5MB/s]
model.safetensors: 100%|███████████████████████████████████████████████████████████████████████████████| 1.81G/1.81G [00:58<00:00, 30.8MB/s]
Fetching 12 files: 100%|████████████████████████████████████████████████████████████████████████████████████| 12/12 [00:59<00:00, 4.99s/it]
Fetching 2 files: 100%|████████████████████████████████████████████████████████████████████████████████████| 2/2 [00:00<00:00, 38479.85it/s]
Error loading model: ModelConfig.init() missing 11 required positional arguments: 'hidden_size', 'num_hidden_layers', 'intermediate_size', 'num_attention_heads', 'rms_norm_eps', 'vocab_size', 'num_key_value_heads', 'max_position_embeddings', 'rope_theta', 'head_dim', and 'tie_word_embeddings'
Traceback (most recent call last):
File "/Users/tulas/Projects/l2p/.venv/lib/python3.12/site-packages/mlx_audio/tts/generate.py", line 262, in generate_audio
model = load_model(model_path=model)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/tulas/Projects/l2p/.venv/lib/python3.12/site-packages/mlx_audio/tts/utils.py", line 249, in load_model
model_class.ModelConfig.from_dict(config)
File "/Users/tulas/Projects/l2p/.venv/lib/python3.12/site-packages/mlx_lm/models/base.py", line 15, in from_dict
return cls(
^^^^
TypeError: ModelConfig.init() missing 11 required positional arguments: 'hidden_size', 'num_hidden_layers', 'intermediate_size', 'num_attention_heads', 'rms_norm_eps', 'vocab_size', 'num_key_value_heads', 'max_position_embeddings', 'rope_theta', 'head_dim', and 'tie_word_embeddings'