xorslider/FoodExtract-gemma-3-270m-finetuned

FoodExtract: Fine-Tuned Gemma 3 270M

This model, xorslider/FoodExtract-gemma-3-270m-finetuned, is a specialized version of Google's Gemma 3 270M, fine-tuned for the precise task of extracting food and drink items from arbitrary text. With 0.3 billion parameters and a context length of 32768 tokens, it's designed for efficient and focused information retrieval.

Key Capabilities

• Structured Extraction: Outputs identified food and drink items in a structured format, along with relevant tags (e.g., nutrition_panel, ingredient_list, menu, recipe, food_items, drink_items, food_advertisement, food_packaging).
• Food/Drink Detection: Accurately identifies the presence or absence of food/drink items in text.
• Compact and Efficient: Built on the Gemma 3 270M base, offering a lightweight solution for specific extraction tasks.

Good For

• Automated Data Processing: Ideal for applications requiring automatic identification and categorization of food and drink mentions from large volumes of text.
• Recipe Analysis: Extracting ingredients from recipes or cooking instructions.
• Menu Digitization: Converting unstructured menu text into structured food item lists.
• Nutritional Information Systems: Parsing food-related text for relevant items to integrate into nutritional databases.
• Content Tagging: Automatically tagging articles, reviews, or advertisements based on food and drink content.