Videos
------

- Active Tuning - Introduction and MSO.mp4
    A short introduction to Active Tuning including some results
    (here with an Echo State Network instead of an LSTM).

- Active Tuning - Wave (0.0).mp4
    Demonstration of Active Tuning used on a fully noise unware network (0.0).

- Active Tuning - Wave (0.05).mp4
    Demonstration of Active Tuning using a network training with a small portion of noise (0.05).



Code (trained models included)
------------------------------

The scripts were tested successfully with Python 3.3.7 with PyTorch 1.4 and PyTorch 1.6.

The following list contains the most relevant source code files. * = holds for all three experiments (mso, pendulum, and wave).

- active_tuning/active_tuning.py
    Generic implementation of Active Tuning that was used in all experiments.


- experiments/*/data_generation.py
    Generates the data used for training and testing.

- experiments/*/launch_training.py
    Launches the training of all models.

- experiments/*/launch_evaluation.py
    Launches the systematic evaluation of all denoiser networks.

- experiments/*/evaluate_model.py
    Particularly used for the individual active tuning evaluations.

- experiments/*/test_model.py
    Mosty used for invidiual experiments and plotting.
