3adel25/leftover-recipe-generator-qwen1.5b
The 3adel25/leftover-recipe-generator-qwen1.5b is a 1.5 billion parameter instruction-tuned Small Language Model (SLM) based on Qwen2.5-1.5B-Instruct, developed by Adel Hani Mohamed Adlan. Fine-tuned using QLoRA, it specializes in generating step-by-step cooking recipes from a list of available fridge ingredients, focusing on practical and creative meal ideas without requiring external ingredients. Its compact size and efficient fine-tuning make it suitable for edge or local deployment on low-VRAM GPUs, CPUs, or mobile devices, primarily serving home cooking, meal planning, and food waste reduction.
Loading preview...
Fridge Leftover Recipe Generator (Qwen2.5-1.5B-Instruct)
This model, developed by Adel Hani Mohamed Adlan, is an instruction-tuned Small Language Model (SLM) built upon the Qwen/Qwen2.5-1.5B-Instruct base. It has been specifically fine-tuned using Parameter-Efficient Fine-Tuning (QLoRA) to excel at generating customized, step-by-step cooking recipes from a given list of leftover fridge ingredients. The model aims to provide fast, practical, and creative meal ideas, emphasizing the use of existing ingredients to reduce food waste.
Key Capabilities
- Recipe Generation: Transforms a list of random fridge ingredients into a complete recipe, including title, ingredient breakdown, step-by-step instructions, and chef tips.
- Food Waste Reduction: Encourages the creative utilization of household ingredients before they expire.
- Efficient Deployment: Its compact 1.5 billion parameter size, combined with QLoRA fine-tuning, allows for efficient operation on low-VRAM GPUs, CPUs, or mobile devices, making it ideal for local or edge deployments.
Good for
- Home cooks and meal planners looking for quick, ingredient-specific recipe ideas.
- Individuals aiming to minimize food waste by creatively using leftovers.
- Developers seeking a specialized, lightweight language model for on-device or resource-constrained applications related to culinary assistance.