"Fetching files" on every generate run?

[rosmur]

Seeing the following everytime I run generate:

None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
Fetching 11 files: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 11/11 [00:00<00:00, 59378.82it/s]
Fetching 11 files: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 11/11 [00:00<00:00, 94157.84it/s]

What is being fetched?

Code:

# Load the model
model_path = "mlx-community/pixtral-12b-8bit"
model, processor = load(model_path)
config = load_config(model_path)

# Apply chat template
formatted_prompt = apply_chat_template(processor, config, prompt, num_images=len(image))

# Generate output
output = generate(
    model, processor, image, formatted_prompt, verbose=False, max_tokens=1000
)
print(output)

[Blaizzy]

The model weights are being loaded :)