C2AE

This is the Tensorflow implementation for the paper 'Learning Deep Latent Spaces for Multi-Label Classfications' published in AAAI 2017.

Installation

The model was built and tested using Python 3! Install the following dependencies :

pip3 install liac-arff 

Tensorflow

Running

This code supports the .arff data format, however if you wish to use any other data format, convert it into numpy arrays and dump it to the data/dataset_name with the name format as mentioned in data/README.md and modify model/src/parser.py.

cd ./model/src
python3 __main__.py

Logs

All the logs are saved in ./model/stdout and you can visualize the loss using tensorboard by pointing it to ./model/results/tensorboard.