jmarianodc/health_food_demo
The jmarianodc/health_food_demo is a 1.1 billion parameter Llama-based language model, fine-tuned for classifying foods as either healthy or unhealthy. This model, available in GGUF format, is specifically designed for educational and demonstration purposes in food classification. It operates with a 2048-token context length and is optimized for local inference tools like LM Studio and llama.cpp.
Loading preview...
Model Overview
The jmarianodc/health_food_demo is a lightweight, 1.1 billion parameter Llama-based language model. It has been specifically fine-tuned using supervised learning to classify food items as either Healthy or Unhealthy. The model is provided in GGUF format, making it highly compatible with various local inference tools.
Key Capabilities
- Food Classification: Determines if a given food is healthy or unhealthy based on its training.
- Local Inference: Optimized for deployment and inference on local machines using tools such as LM Studio, llama.cpp, Ollama (with conversion), KoboldCpp, and Text Generation WebUI.
- GGUF Format: Available in quantized (Q4_K_M) and full precision GGUF formats for flexible use depending on memory constraints and desired quality.
- Educational/Demo Focus: Primarily intended for learning, demonstration, and experimentation in the domain of food classification.
Training Details
The model was trained on a custom food classification dataset, which included examples of both healthy and unhealthy foods. The training format involved question-answer pairs, such as "Is Apples healthy or unhealthy?" followed by the answer "Healthy".
Important Considerations
- Not for Medical Advice: This model is a small educational/demo model and should not be used for medical or nutritional advice.
- Quantization Impact: The quantized versions, while memory-efficient, may yield slightly different results compared to the full precision model.