Parallel trainers

This package trains a parallel version of the following models, using the python binds for MPI (mpi4py - https://pypi.python.org/pypi/mpi4py/), and the Trainers and Machines from Bob:

• Universal Background Model (UBM)
• The within client variation matrix (U Matrix) for the Intersession Variability Modeling (ISV)
• Total Variability Matrix (T Matrix)
• Train Linear Projectors for iVectors (Whitening, LDA and WCCN)
• IVectors generation

With these codes you can split all the work in a grid system that has MPI available.

If you use this package, please cite the following publications:

1. The original paper for the UBM system:

   @article{reynolds2000speaker,
     title={Speaker verification using adapted Gaussian mixture models},
     author={Reynolds, Douglas A and Quatieri, Thomas F and Dunn, Robert B},
     journal={Digital signal processing},
     volume={10},
     number={1},
     pages={19--41},
     year={2000},
     publisher={Elsevier}
   }
   

2. Paper describing the use of Session Variability Modelling for face authentication:

   @article{mccool2013session,
     title={Session variability modelling for face authentication},
     author={McCool, Christopher and Wallace, Roy and McLaren, Mitchell and El Shafey, Laurent and Marcel, S{\'e}bastien},
     journal={IET biometrics},
     volume={2},
     number={3},
     pages={117--129},
     year={2013},
     publisher={IET}
   }
   

3. Bob as the core framework used to train the models:

   @inproceedings{Anjos_ACMMM_2012,
     author = {A. Anjos AND L. El Shafey AND R. Wallace AND M. G\"unther AND C. McCool AND S. Marcel},
     title = {Bob: a free signal processing and machine learning toolbox for researchers},
     year = {2012},
     month = oct,
     booktitle = {20th ACM Conference on Multimedia Systems (ACMMM), Nara, Japan},
     publisher = {ACM Press},
   }
   

Installation

Using zc.buildout

Download the latest version of this package from Github and unpack it in your working area. The installation of the toolkit itself uses buildout. You don't need to understand its inner workings to use this package. Here is a recipe to get you started:

$ python bootstrap.py
$ ./bin/buildout

These 2 commands should download and install all non-installed dependencies and get you a fully operational test and development environment.

Note

The python shell used in the first line of the previous command set determines the python interpreter that will be used for all scripts developed inside this package. Because this package makes use of Bob, you must make sure that the bootstrap.py script is called with the same interpreter used to build Bob, or unexpected problems might occur.

If Bob is installed by the administrator of your system, it is safe to consider it uses the default python interpreter. In this case, the above 3 command lines should work as expected. If you have Bob installed somewhere else on a private directory, edit the file buildout.cfg before running ./bin/buildout. Find the section named external and edit the line egg-directories to point to the lib directory of the Bob installation you want to use. For example:

[external]
recipe = xbob.buildout:external
egg-directories=/Users/crazyfox/work/bob/build/lib

User Guide

Universal Background Training

Type the following command to see all the available options for the UBM trainer:

$ ./bin/ubm_trainer.py --help

In order to run this script in the MPI environment run the following code:

$ mpirun --np <number_of_nodes> --hosts=<available_hosts (comma separated)> ./bin/ubm_trainer.py <options>

It is possible to use an xbob.db package as input or a file containing the list of features to train. To use a database, run the following code:

$ mpirun --np <number_of_nodes> --hosts=<available_hosts (comma separated)> ./bin/ubm_trainer.py <options> database -d <database_name>

To use a regular file list, run the following code::

$ mpirun --np <number_of_nodes> --hosts=<available_hosts (comma separated)> ./bin/ubm_trainer.py <options> list -f <file_name>

Within client variation matrix (U Matrix) for the ISV

Type the following command to see all the available options for the ISV (U Matrix) trainer:

$ ./bin/isv_U_trainer.py --help

In order to run this script in the MPI environment run the following code:

$ mpirun --np <number_of_nodes> --hosts=<available_hosts (comma separated)> ./bin/isv_U_trainer.py <options>

Total Variability Matrix (T Matrix)

Type the following command to see all the available options for the Total Variability Matrix trainer:

$ ./bin/ivector_TV_trainer.py --help

In order to run this script in the MPI environment run the following code:

$ mpirun --np <number_of_nodes> --hosts=<available_hosts (comma separated)> ./bin/ivector_TV_trainer.py <options>

Train Linear Projectors for iVectors (Whitening, LDA and WCCN)

Type the following command to see all the available options to train Linear projectors for iVectors:

$ ./bin/train_linear_machines.py --help

Unfortunately there is no parallel implementation for this code.

iVectors generation

Type the following command to see all the available options for iVectors generation:

$ ./bin/generate_ivectors.py --help

In order to run this script in the MPI environment run the following code:

$ mpirun --np <number_of_nodes> --hosts=<available_hosts (comma separated)> ./bin/generate_ivectors.py <options>

How to configure the MPI in my grid system?

You can see all the details of how to configure the MPI and how to setup the python bindings in the following page: http://mpi4py.scipy.org/.

Problems

In case of problems, please contact any of the authors of the package.