How prince-khatri/coffee_roast works
Project type and purpose
• This is a small Python-based machine-learning pipeline for classifying coffee roasts.
• It provides two main scripts: one to train a TensorFlow/Keras model (trainModel.py), and one to run inference (inference.py).
• There is also a utility module for data loading and normalization (coffeeRoastUtils.py) and a GitHub Actions CI/CD workflow.
File structure
Top level:
• trainModel.py → orchestrates data loading, model definition, training, and model serialization
• inference.py → loads saved model and runs predictions on new inputs
• coffeeRoastUtils.py → contains dataset loading, normalization, and helper functions
• requirements.txt → Python dependencies (TensorFlow, Keras, etc.)
• .github/workflows/ → CI pipeline YAML that installs deps, trains the model, and runs inference on each push
README for additional insights
• Training and inference commands are shell‐based scripts.
• CI workflow automates install → train → predict on GitHub push.
• Data sources are assumed to be local or specified within the utils module.
• The model is a simple feedforward NN for binary classification.