Tutorials
==========

Runnable example notebooks live in the GitHub repository:

- `01_Beginner_Easy_Start.ipynb
  <https://github.com/KagakuLab/QSARmil/blob/main/notebooks/01_Beginner_Easy_Start.ipynb>`_
  — the predefined, zero-configuration pipeline: SMILES and targets in,
  predictions out, with no separate save/load step.
- `02_Professional_Pipeline_Customization.ipynb
  <https://github.com/KagakuLab/QSARmil/blob/main/notebooks/02_Professional_Pipeline_Customization.ipynb>`_
  — building a custom pipeline from individual modules (conformer
  generator, descriptor calculator, a MIL network, hyperparameter search).
- `03_Key_Instance_Detection.ipynb
  <https://github.com/KagakuLab/QSARmil/blob/main/notebooks/03_Key_Instance_Detection.ipynb>`_
  — identifying which conformer drives a prediction, and checking those
  guesses against a benchmark with known correct answers.