This is the Tensorflow implementation for the paper 'Learning Deep Latent Spaces for Multi-Label Classfications' published in AAAI 2017.
The model was built and tested using Python 3! Install the following dependencies :
pip3 install liac-arff
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
All the logs are saved in ./model/stdout
and you can visualize the loss using tensorboard by pointing it to ./model/results/tensorboard
.