Joint Detection-Classification Model
This is the code for "A Joint Detection-Classification Model for Audio Tagging of Weakly Labelled Data", https://arxiv.org/pdf/1610.01797.pdf
02/10/2016
=====================================

pre-requisite python package:
* librosa	(feature extraction)
* hat		(deep learning toolbox)

Usage:
* modify config.py, config your own paths

Development
* run prepare_dev_data.py
* run main_dev_bob.py (or main_dev_jdc.py), (you need to select the fold between 0 - 4)
* run calculate_results.py (You need to select one to run in line 59 - 62)

Evaluation
* run prepare_eva_data.py
* run main_eva_bob.py (or main_eva_jdc.py)
* run calculate_results.py (You need to select one to run in line 59 - 62)

Plot
* run plot_results.py (must run after running all folds for main_dev_bob.py, main_dev_jdc.py and main_eva_bob.py, main_eva_jdc.py)