Team Members: Pascal Reich, Tammy Dahl, Jordan Isaiah Espiritu, Evan Tardiff, Dylan Murphy, Jwann Sy
Our AI Project is an AI Generated Recommendations for Recipes. Its purpose is to provide a program that can suggests recipes for users based on the ingredients or the type of recipe that is being searched. Based on the selections, the program will use AI algorithms, such as Word2Vec, K-Mean Clusters, and Cosine Similarity to compare similarity with other recipes and provide more personalized suggestions
This program requires Python 3.10 or later
Install the necessary libraries using pip
pip install numpy pandas scikit-learn gensim gradio kaggle
Download the RecipeNLG Dataset from Kaggle
- Obtain a Kaggle API token (
kaggle.json) and set it up
mkdir ~/.kaggle
cp path_to_kaggle.json ~/.kaggle/
chmod 600 ~/.kaggle/kaggle.json
- Download the Dataset
kaggle datasets download -d paultimothymooney/recipenlg
- Extract the Dataset
unzip recipenlg.zip
Run the project effortlessly using Google Colab. Click the button in .ipynb file to launch the notebook
You can also run the project locally:
-
Clone the repository and navigate to the project folder.
-
Start a Jupyter notebook or execute scripts directly in your IDE.
⚠ Warning: Local execution may require significant storage, memory, and computational resources due to the size of the dataset and the training requirements.
-
Ingredient-Based Recommendations: Enter your favorite ingredients to discover matching recipes.
-
Recipe ID Recommendations: Use a specific recipe ID to find similar dishes.
-
Cluster Recommendations: Explore recipes within the same category.
Train the recommendation system using the following steps:
-
Load and preprocess the RecipeNLG dataset.
-
Train a Word2Vec model on the cleaned ingredient lists.
-
Generate recipe vectors and cluster them using K-means.
-
Save the trained models for inference.
Run the CLI cell to interact with the system via command line.
Launch the Gradio cell for a user-friendly experience.