Ingredients_finder

🛒 Shopping List → Recipe Generator (Flan-T5) 📌 Project Overview

This project uses the Flan-T5 model from Hugging Face to generate recipe ideas from a given shopping list. It is a text-to-text generation mini-project in Generative AI.

Input: A shopping list or paragraph containing ingredients.

Output: Creative recipe names with short descriptions.

Example:

Input: tomatoes, pasta, cheese, chicken Output:

  1. Cheesy Chicken Pasta Bake – Pasta baked with tomato sauce, chicken, and cheese.
  2. Grilled Chicken Spaghetti – Spaghetti in fresh tomato sauce topped with grilled chicken.
  3. Creamy Tomato Chicken Lasagna – Layers of pasta, chicken, cheese, and tomato cream sauce.

🛠️ Tech Stack

Python

Hugging Face Transformers

Flan-T5 (google/flan-t5-small or google/flan-t5-base)

(Optional) Streamlit – for a simple web UI

🚀 Installation

Clone this repo (or copy the code files):

git clone https://github.com/your-username/recipe-generator.git cd recipe-generator

Install dependencies:

pip install transformers sentencepiece streamlit

<img widthScreenshot 2025-09-04 105432 =”1085” height=”189” alt=”Screenshot 2025-09-04 105409” src=”https://github.com/user-attachments/assets/e68dcfd6-90e4-4233-8121-7017d5b3cbc2” />

Run the Python script:

python recipe_generator.py

(Optional) Run the Streamlit app:

streamlit run app.py

đź“„ Usage Python Script from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

model_name = “google/flan-t5-base” tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForSeq2SeqLM.from_pretrained(model_name)

def generate_recipe(shopping_list): prompt = ( f”You are a professional chef. Suggest 3 creative recipes using these ingredients: {shopping_list}. “ “Each recipe should include a dish name and a short description.” ) inputs = tokenizer(prompt, return_tensors=”pt”, max_length=512, truncation=True) outputs = model.generate(**inputs, max_length=220, num_beams=5, no_repeat_ngram_size=2, early_stopping=True) return tokenizer.decode(outputs[0], skip_special_tokens=True)

print(generate_recipe(“bread, butter, milk, honey”))

Streamlit App (Optional UI) import streamlit as st

st.title(“🛒 Shopping List → Recipe Generator”)

user_input = st.text_area(“Enter your shopping list:”)

if st.button(“Generate Recipes”): if user_input.strip(): recipes = generate_recipe(user_input) st.write(“### 🍴 Suggested Recipes:”) st.write(recipes) else: st.warning(“Please enter some ingredients!”)

đź“Š Features

âś… Generates multiple recipe ideas from ingredients âś… Works with simple lists or full paragraphs with scattered ingredients âś… Can be extended with recipe categories (breakfast, dinner, dessert, etc.) âś… Optional web app with Streamlit for easy testing

đź”® Future Improvements

Fine-tune the model on real recipe datasets

Add cuisine-specific outputs (Indian, Italian, Vegan, etc.)

Generate detailed step-by-step cooking instructions