Multi-label Classification with Deep Learning

This is a Python3.5 version and keras 2.2.0 implementation for the paper: ``ADIOS: Architectures Deep In Output Space"

adios.utils.assemble.assemble helper function provides and handy way to construct ADIOS and MLP models from config dictionaries.

All example scripts are given in scripts/.

Note: keras.models.Graph was no longer supported starting from keras-v1.2. The last version of ADIOS used the legacy code, keras.legacy.models.Graph, but it was removed in next versions. Thus we replaced it with Model from the Functional API.

Requirements

• scikit-multilearn
• NumPy
• pyyaml
• Theano==1.0.2
• keras==2.2.0
• scikit-learn

The requirements can be installed via pip as follows:

$ pip install -r requirements.txt

Optional (needed only for using Jobman):

• argparse
• Jobman

Installation

To use the code, we recommend installing it as Python package in the development mode as follows:

$ python setup.py develop [--user]

The --user flag (optional) will install the package for a given user only.

This version shows an example of the Delicious dataset, following the MULAN format. You may also need to switch the scikit-multilearn arff reader to non-binary, so that you parse the datasets.

Example

To run the ADIOS model with MBC, edit the scripts/adios_train.py change labels_order to 'delicious'. Then for the GPU version run:

$ THEANO_FLAGS=device=cuda0,floatX=float32 python3.5 adios_train.py