juanpajedrez/FoodExtract-gemma-3-270m-fine-tune-v1-VIDEO
The juanpajedrez/FoodExtract-gemma-3-270m-fine-tune-v1-VIDEO model is a Gemma 3 270M-based language model fine-tuned for food and drink extraction from raw text. It classifies text as food/drink, tags it with predefined categories, and extracts edible food and drink items into separate lists. This model is optimized for filtering large image caption datasets to identify food and drink related content.
Loading preview...
Overview
This model, juanpajedrez/FoodExtract-gemma-3-270m-fine-tune-v1-VIDEO, is a specialized language model built upon the Gemma 3 270M architecture. Its core function is to process raw text and perform structured extraction of food and drink related information. It was fine-tuned using Supervised Fine-Tuning (SFT) with Hugging Face's TRL library on the FoodExtract-1k dataset, which contains 1400 samples of raw text and corresponding structured food extractions generated by gpt-oss-120b.
Key Capabilities
- Food/Drink Classification: Determines if input text contains food or drink references.
- Tagging: Assigns one or more predefined tags (e.g.,
nutrition_panel,ingredient list,food_items,drink_items) for fast filtering. - Item Extraction: Identifies and lists specific edible food and drink items mentioned in the text.
- Condensed Output: Generates a streamlined output format to reduce token generation, with helper functions provided for condensing and uncondensing the structured data.
Good For
- Dataset Filtering: Ideal for filtering large text datasets, such as image captions (e.g., DataComp-1B), to isolate food and drink related entries.
- Structured Data Generation: Transforming unstructured text into structured food and drink data for further analysis or application.