LibriLight ABX

This is a wrapper module around the abx implementation found in libri-light/eval. This module only adds a wrapper function to directly call abx evaluation and a dataclass object annotating all arguments as well as a setup.py to allow installation as a module.

Installation

You can install this module from pip directly using the following command : pip install zerospeech-libriabx

Or you can install from source by cloning this repository and running : pip install .

Usage

From command line

A command line is created to allow running abx evaluations.

usage: libri-abx [-h] [--path_checkpoint PATH_CHECKPOINT] [--file_extension {.pt,.npy,.wav,.flac,.mp3}] [--feature_size FEATURE_SIZE] [--cuda] [--mode {all,within,across}]
                 [--distance_mode {euclidian,cosine,kl,kl_symmetric}] [--max_size_group MAX_SIZE_GROUP] [--max_x_across MAX_X_ACROSS] [--out OUT]
                 path_data path_item_file

ABX metric

positional arguments:
  path_data             Path to directory containing the data
  path_item_file        Path to the .item file

optional arguments:
  -h, --help            show this help message and exit
  --path_checkpoint PATH_CHECKPOINT
                        Path to a CPC checkpoint. If set, the apply the model to the input data to compute the features
  --file_extension {.pt,.npy,.wav,.flac,.mp3}
  --feature_size FEATURE_SIZE
                        Size (in s) of one feature
  --cuda                Use the GPU to compute distances
  --mode {all,within,across}
                        Choose the mode of the ABX score to compute
  --distance_mode {euclidian,cosine,kl,kl_symmetric}
                        Choose the kind of distance to use to compute the ABX score.
  --max_size_group MAX_SIZE_GROUP
                        Max size of a group while computing theABX score. A small value will make the code faster but less precise.
  --max_x_across MAX_X_ACROSS
                        When computing the ABX across score, maximumnumber of speaker X to sample per couple A,B. A small value will make the code faster but less precise.
  --out OUT             Path where the results should be saved

From python API

To call the abx evaluation from python code you can use the following example :

from pathlib import Path
import libriabx

args = libriabx.AbxArguments(
    path_data=Path("/location/to/scores/")
    path_item_file=Path("/location/to/file.item")
    **other_options
)
result = libriabx.abx_eval(args)

For all possible options see the AbxArguments class definition.

Result is a dictionary containing Dict{mode -> score} where mode is defined as (across, within)

Building & Upload

For security & compatibility reasons binary builds require a special environment to be build. We need to use the manylinux docker container so that correct flags are used.

To do this run :

> docker pull quay.io/pypa/manylinux2014_x86_64
> docker run --rm -v `pwd`:/io quay.io/pypa/manylinux2014_x86_64 bash /io/build_wheel.sh

To check for pyversions available docker run --rm quay.io/pypa/manylinux2014_x86_64 ls /opt/python

This allows to populate with compiled versions of libriabx for python3.8, python3.9, python3.10, python3.11 in the dist folder. For more information see ManyLinux Implementation

Once binaries have been build we can upload them to pypi using : twine upload dist/*