Machine Learning Template

This project provides a machine learning quickstart template using PyTorch Lightning, Hydra, MLflow, and uv.

Installation

Clone the repository:

git clone <your-repository-url>
cd machine-learning

Install uv: Follow the instructions on the uv website if you don't have it installed.
Install dependencies: This command creates a virtual environment and installs all necessary packages defined in pyproject.toml.
```
uv sync
```

This project uses Hydra for configuration management.

Configuration files are located in the configs/ directory.
The main configuration file is configs/lymph_nodes.yaml.
You can override configuration parameters directly from the command line. For example, to change the batch size:
```
uv run python +m <lymph_nodes> mode=fit data.batch_size=64
```
MLflow is configured as the default logger (see configs/default.yaml). Ensure your MLflow tracking server is running or configure it accordingly.

Train the model:
```
uv run python +m <lymph_nodes> mode=fit
```
Validate the model: Requires a checkpoint path to be set in the configuration (e.g., checkpoint=path/to/your/checkpoint.ckpt) or passed via the command line.
```
uv run python +m <lymph_nodes> mode=validate checkpoint=path/to/checkpoint.ckpt
```

Test the model: Requires a checkpoint path.

uv run python +m <lymph_nodes> mode=test checkpoint=path/to/checkpoint.ckpt

Run prediction: Requires a checkpoint path.

uv run python +m <lymph_nodes> mode=predict checkpoint=path/to/checkpoint.ckpt

uvx ruff check  # Check and fix linting issues
uvx ruff format # Format code
uvx mypy .     # Run mypy