ML Model Factory 🏭

An automated, end-to-end machine learning pipeline for training, evaluating, and versioning models. This project was built to demonstrate a foundational understanding of MLOps principles.

---

Project Overview

This pipeline automates the entire machine learning workflow. It starts by ingesting raw data, proceeds to clean and preprocess it, trains a machine learning model, tunes its hyperparameters, and finally versions the resulting model and its performance metrics.

---

Features

• Automated Data Preprocessing: Handles missing values, scales numerical features, and encodes categorical features.
• Model Training: Trains a classification/regression model on the prepared data.
• Hyperparameter Tuning: Automatically finds the best hyperparameters using techniques like Grid Search.
• Model Versioning: Uses MLflow to track experiments, log metrics, and version models.

---

Tech Stack

• Language: Python
• Data Handling: Pandas, NumPy
• ML Framework: Scikit-learn
• Experiment Tracking: MLflow
• Orchestration: Python scripts

---

How to Run

1. Clone the repository:

   git clone https://github.com/Fugant1/ml-model-factory.git

2. Install the dependencies:

   pip install -r requirements.txt

3. Move your dataset to the data folder:

   mv your_dataset_file data/

4. Configure the pipeline: Open src/config.py and set the FILENAME, TARGET_COLUMN, and MODEL_SELECTED variables to match your data and preferences.

5. Run the main pipeline script:

   python -m src.pipeline.py