Water Cherenkov Machine Learning Variational AutoEncoder (WatChMaL-VAE)

Description

Python implementation of the training engine and framework to build, train and test VAE models for Water Cherenkov Detectors.

1. Directory Layout

2. Installation

3. Usage

4. Credits

5. License

Directory Layout

.
+-- config                              # Configuration files
  +-- engine_config                     # Configuration files for the training engine
    +-- test_resnet.ini
    +-- test_kazunet.ini
    +-- test_kvaenet.ini
  +-- preprocessing_config              # Configuration files for data pre-processing
    +-- merge_config.ini
+-- io_utils                            # Tools to handle the user inputs, dataset and models
  +-- __init__.py
  +-- arghandler.py
  +-- data_handling.py
  +-- ioconfig.py
  +-- modelhandler.py
+-- models                              # PyTorch implementation of various CNN-VAE models
  +-- abhinet.py
  +-- convonlynet.py
  +-- densenet.py
  +-- resnet.py
  +-- kazunet.py
  +-- kvaenet.py
  +-- resnet.py
  +-- vaenet.py
+-- notebooks                           # Jupyter Notebooks for offline analysis
+-- plot_utils                          # Tools for visualizing model performance and dataset features
  +-- mpmt_visual.py
  +-- notebook_utils.py
  +-- plot_utils.py
+-- postprocessing                      # Tools for post-processing the outputs from the models
+-- preprocessing                       # Tools for pre-processing the dataset
  +-- merge_h5.py
  +-- merge_numpy_arrays_hdf5.py 
  +-- preprocessing_gamma.py
+-- root_utils                          # Tools for interacting with the ROOT files from the WCSim simulations
  +-- display_list.py
  +-- event_disp_and_dump.py
  +-- event_disp_and_dump_arg_utils.py
  +-- event_display.py
  +-- event_dump.py
  +-- event_dump_cedar_start.py
  +-- event_dump_one.py
  +-- pos_utils.py
+-- training_utils                      # Tools for training, validating and testing the models
  +-- doublepriorityqueue.py
  +-- engine.py
  +-- engine_vae.py
+-- README.md                           # README documentation for the repository
+-- USER_GUIDE.md                       # User guide on how to use the library
+-- watchmal.py                         # Main script to run the engine for the classifier
+-- watchmal_vae.py                     # Main script to run the engine for the VAE

Installation

Requirements

The following Python standard, machine learning and deep learning libraries are required for the functionality of the framework :

To download the repository use :

git clone https://github.com/WatChMaL/VAE.git

Usage

Sample configuration files for the training engine and framework are provided in config/engine_config.

Examples :

# Train and validate a classifier using the sample configuration file

python3 watchmal.py -l test_resnet

# Train and validate a variational autoencoder using the sample configuration file

python3 watchmal_vae.py -l test_kvaenet

More in-depth usage options and configurations are described in USER_GUIDE.md.

ParasKoundal/VAE