An enhancement to a machine learning model that predicts whether a recipe is for a muffin or a cupcake, now dressed with a user-friendly web interface. Built using NumPy, Pandas, Scikit-Learn, FastAPI, HTML, CSS, and JavaScript.
This project features:
- A trained Random Forest Classifier that distinguishes between muffin and cupcake recipes.
- A simple web interface powered by FastAPI and Jinja2 templates.
- An interactive API endpoint to classify new recipes based on ingredients.
- A
/model-inforoute to inspect the internal model configuration.
| Category | Tools / Libraries |
|---|---|
| Programming Lang | Python 3, HTML, CSS, JavaScript |
| Data | NumPy, Pandas |
| Machine Learning | Scikit-learn (RandomForestClassifier) |
| Web Framework | FastAPI |
| Templating | Jinja2 |
| Packaging | joblib |
├── ml_components/ │ ├── muffins_vs_cupcakes.csv # Training dataset │ └── model.joblib # Saved ML model ├── templates/ │ └── base.html # UI Template ├── main.py # FastAPI app ├── README.md # Project documentation
The dataset (muffins_vs_cupcakes.csv) contains ingredient quantities and a label identifying whether the recipe is a muffin or a cupcake.
Features:
- Flour, Milk, Sugar, Butter, Egg, Baking Powder, Vanilla, Salt
Target:
Type:MuffinorCupcake(encoded as 1 and 0)
-
Training the Model
ARandomForestClassifieris trained on the labeled recipe dataset and saved usingjoblib. -
Serving Predictions
The FastAPI app loads the trained model and serves predictions via:POST /predict: Accepts ingredient values and returns a prediction.GET /model-info: Returns model parameters.
-
User Interface
A lightweight HTML page allows manual input and displays the predicted recipe type.
POST /predict
{
"Flour": 2,
"Milk": 1,
"Sugar": 1,
"Butter": 1,
"Egg": 1,
"Baking_Powder": 1,
"Vanilla": 1,
"Salt": 1
}
{
"prediction": "Cupcake"
}
pip install -r requirements.txt
uvicorn main:app --reload