gutsignal-food-parser-llama-3.2-1b

GutSignal Food Parsing Model - MLX Format

Fine-tuned language model for parsing food descriptions into structured data, optimized for Apple Silicon.

Model Description

This model is fine-tuned from meta-llama/Llama-3.2-1B to parse natural language food descriptions and extract:

• Individual ingredients
• Food categories (dairy, grains, protein, vegetables, etc.)
• Beverage classification
• Dairy content detection
• Estimated nutritional information

Format

This model is in Apple MLX format, optimized for on-device inference on Apple Silicon (M1/M2/M3/M4 chips) and iOS devices.

Intended Use

• Primary Use: On-device food parsing for health tracking applications
• Target Platform: iOS devices (iPhone 15+), macOS (Apple Silicon)
• Privacy: Designed for offline, on-device inference

Usage

On macOS (MLX)

from mlx_lm import load, generate

model, tokenizer = load("ndlanier/gutsignal-food-parser-llama-3.2-1b")

prompt = """### Instruction:
Parse the following food description and extract structured data.

### Input:
chicken salad with ranch dressing

### Output:
"""

response = generate(model, tokenizer, prompt=prompt, max_tokens=256)
print(response)

On iOS

See iOS Integration Guide for using MLX models in iOS apps.

Output Format

{
  "ingredients": ["chicken", "lettuce", "ranch dressing"],
  "categories": ["protein", "vegetables", "fats"],
  "is_beverage": false,
  "contains_dairy": true,
  "estimated_calories": 450
}

Categories

dairy, grains, protein, vegetables, fruits, fats, sugars, beverages, spicy, fiber, processed, caffeine, alcohol, fermented, nuts, legumes, unknown

Training Details

• Base Model: meta-llama/Llama-3.2-1B
• Fine-tuning Method: LoRA (Low-Rank Adaptation)
• Training Data: Food journal entries, USDA FoodData Central, Open Food Facts
• Format: MLX (converted from fine-tuned PyTorch model)

Limitations

• Estimates only - not for medical decisions
• Best performance on common foods
• Approximate calorie estimates

License

llama3.2

Citation

@misc{gutsignal-food-parser-mlx,
  author = {Nathan Lanier},
  title = {GutSignal Food Parsing Model - MLX},
  year = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/ndlanier/gutsignal-food-parser-llama-3.2-1b}}
}