mlx-community/gemma-3-270m-it-4bit · ValueError: Expected shape (262144, 640) but received shape (262144, 80) for parameter lm

Developer % mkdir gemma
Developer % cd gemma 
gemma % py11
(3.11) gemma % pip install mlx-lm
Collecting mlx-lm
  Downloading mlx_lm-0.26.3-py3-none-any.whl.metadata (10 kB)
Collecting mlx>=0.26.0 (from mlx-lm)
  Downloading mlx-0.28.0-cp311-cp311-macosx_15_0_arm64.whl.metadata (5.7 kB)
Requirement already satisfied: numpy in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from mlx-lm) (1.26.4)
Collecting transformers>=4.39.3 (from mlx-lm)
  Downloading transformers-4.55.2-py3-none-any.whl.metadata (41 kB)
Collecting protobuf (from mlx-lm)
  Downloading protobuf-6.32.0-cp39-abi3-macosx_10_9_universal2.whl.metadata (593 bytes)
Requirement already satisfied: pyyaml in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from mlx-lm) (6.0.2)
Requirement already satisfied: jinja2 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from mlx-lm) (3.1.6)
Collecting mlx-metal==0.28.0 (from mlx>=0.26.0->mlx-lm)
  Downloading mlx_metal-0.28.0-py3-none-macosx_15_0_arm64.whl.metadata (5.1 kB)
Collecting filelock (from transformers>=4.39.3->mlx-lm)
  Downloading filelock-3.19.1-py3-none-any.whl.metadata (2.1 kB)
Collecting huggingface-hub<1.0,>=0.34.0 (from transformers>=4.39.3->mlx-lm)
  Using cached huggingface_hub-0.34.4-py3-none-any.whl.metadata (14 kB)
Requirement already satisfied: packaging>=20.0 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from transformers>=4.39.3->mlx-lm) (25.0)
Collecting regex!=2019.12.17 (from transformers>=4.39.3->mlx-lm)
  Using cached regex-2025.7.34-cp311-cp311-macosx_11_0_arm64.whl.metadata (40 kB)
Requirement already satisfied: requests in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from transformers>=4.39.3->mlx-lm) (2.32.4)
Collecting tokenizers<0.22,>=0.21 (from transformers>=4.39.3->mlx-lm)
  Using cached tokenizers-0.21.4-cp39-abi3-macosx_11_0_arm64.whl.metadata (6.7 kB)
Collecting safetensors>=0.4.3 (from transformers>=4.39.3->mlx-lm)
  Using cached safetensors-0.6.2-cp38-abi3-macosx_11_0_arm64.whl.metadata (4.1 kB)
Requirement already satisfied: tqdm>=4.27 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from transformers>=4.39.3->mlx-lm) (4.67.1)
Requirement already satisfied: fsspec>=2023.5.0 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.34.0->transformers>=4.39.3->mlx-lm) (2025.7.0)
Requirement already satisfied: typing-extensions>=3.7.4.3 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from huggingface-hub<1.0,>=0.34.0->transformers>=4.39.3->mlx-lm) (4.14.1)
Collecting hf-xet<2.0.0,>=1.1.3 (from huggingface-hub<1.0,>=0.34.0->transformers>=4.39.3->mlx-lm)
  Using cached hf_xet-1.1.7-cp37-abi3-macosx_11_0_arm64.whl.metadata (703 bytes)
Requirement already satisfied: MarkupSafe>=2.0 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from jinja2->mlx-lm) (3.0.2)
Requirement already satisfied: charset_normalizer<4,>=2 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from requests->transformers>=4.39.3->mlx-lm) (3.4.2)
Requirement already satisfied: idna<4,>=2.5 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from requests->transformers>=4.39.3->mlx-lm) (3.10)
Requirement already satisfied: urllib3<3,>=1.21.1 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from requests->transformers>=4.39.3->mlx-lm) (2.5.0)
Requirement already satisfied: certifi>=2017.4.17 in /Users/aaurelions/.venv/3.11/lib/python3.11/site-packages (from requests->transformers>=4.39.3->mlx-lm) (2025.8.3)
Downloading mlx_lm-0.26.3-py3-none-any.whl (235 kB)
Downloading mlx-0.28.0-cp311-cp311-macosx_15_0_arm64.whl (540 kB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 540.3/540.3 kB 6.5 MB/s  0:00:00
Downloading mlx_metal-0.28.0-py3-none-macosx_15_0_arm64.whl (33.2 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 33.2/33.2 MB 2.2 MB/s  0:00:14
Downloading transformers-4.55.2-py3-none-any.whl (11.3 MB)
   ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.3/11.3 MB 4.6 MB/s  0:00:02
Using cached huggingface_hub-0.34.4-py3-none-any.whl (561 kB)
Using cached hf_xet-1.1.7-cp37-abi3-macosx_11_0_arm64.whl (2.6 MB)
Using cached tokenizers-0.21.4-cp39-abi3-macosx_11_0_arm64.whl (2.7 MB)
Using cached regex-2025.7.34-cp311-cp311-macosx_11_0_arm64.whl (285 kB)
Using cached safetensors-0.6.2-cp38-abi3-macosx_11_0_arm64.whl (432 kB)
Downloading filelock-3.19.1-py3-none-any.whl (15 kB)
Downloading protobuf-6.32.0-cp39-abi3-macosx_10_9_universal2.whl (426 kB)
Installing collected packages: safetensors, regex, protobuf, mlx-metal, hf-xet, filelock, mlx, huggingface-hub, tokenizers, transformers, mlx-lm
Successfully installed filelock-3.19.1 hf-xet-1.1.7 huggingface-hub-0.34.4 mlx-0.28.0 mlx-lm-0.26.3 mlx-metal-0.28.0 protobuf-6.32.0 regex-2025.7.34 safetensors-0.6.2 tokenizers-0.21.4 transformers-4.55.2
(3.11) gemma % nano app.py
(3.11) gemma % python app.py 
added_tokens.json: 100%|█████████████████████| 35.0/35.0 [00:00<00:00, 41.2kB/s]
special_tokens_map.json: 100%|█████████████████| 662/662 [00:00<00:00, 1.19MB/s]
model.safetensors.index.json: 34.9kB [00:00, 27.1MB/s]| 0.00/662 [00:00<?, ?B/s]
generation_config.json: 100%|██████████████████| 173/173 [00:00<00:00, 1.20MB/s]
chat_template.jinja: 1.53kB [00:00, 5.29MB/s]         | 0.00/173 [00:00<?, ?B/s]
config.json: 1.65kB [00:00, 919kB/s]B/s]
tokenizer_config.json: 1.16MB [00:00, 9.66MB/s]  | 3/10 [00:00<00:01,  5.36it/s]
tokenizer.model: 100%|█████████████████████| 4.69M/4.69M [00:02<00:00, 1.63MB/s]
tokenizer.json: 100%|██████████████████████| 33.4M/33.4M [00:08<00:00, 3.96MB/s]
model.safetensors: 100%|█████████████████████| 151M/151M [00:23<00:00, 6.41MB/s]
Fetching 10 files: 100%|████████████████████████| 10/10 [00:24<00:00,  2.41s/it]
Traceback (most recent call last):
  File "/Users/aaurelions/Developer/gemma/app.py", line 3, in <module>
    model, tokenizer = load("mlx-community/gemma-3-270m-it-4bit")
                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/aaurelions/.venv/3.11/lib/python3.11/site-packages/mlx_lm/utils.py", line 254, in load
    model, config = load_model(model_path, lazy)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/aaurelions/.venv/3.11/lib/python3.11/site-packages/mlx_lm/utils.py", line 215, in load_model
    model.load_weights(list(weights.items()), strict=strict)
  File "/Users/aaurelions/.venv/3.11/lib/python3.11/site-packages/mlx/nn/layers/base.py", line 200, in load_weights
    raise ValueError(
ValueError: Expected shape (262144, 640) but received shape (262144, 80) for parameter lm_head.weight