Skip to content
Snippets Groups Projects
Commit 86f0001e authored by Saeed SARFJOO's avatar Saeed SARFJOO
Browse files

link audio extractor doc

parent 8d1f8177
Branches
Tags
1 merge request!19Add audio extractor
......@@ -2,14 +2,11 @@
User guide
==========
This package has been done with the intent of using it in conjunction with
both ``bob.learn.pytorch`` to run speaker verification experiments.
In particular, the goal here is to extract features from an input audio using CNN models
trained with PyTorch.
This subpackage is part of ``bob.learn.pytorch`` package to extract features from an input audio using CNN models which
trained with pytorch_.
For this purpose, you can specify your feature extractor in configuration
file to be used together with the ``verifiy.py`` script from ``bob.bio.base``.
file to be used together with the ``verifiy.py`` script from :ref:`bob.bio.base <bob.bio.base>`.
DltResNet Model
......@@ -50,7 +47,7 @@ A concrete example
Imagine that you have the DltResNet model and you would
like to use the embedding layer as a feature to encode identity.
Your `preprocessor` in Bob pipe-lines should be defined this way in the configuration file:
Your ``preprocessor`` in bob_ pipe-lines should be defined this way in the configuration file:
.. code:: python
......@@ -61,7 +58,7 @@ Your `preprocessor` in Bob pipe-lines should be defined this way in the configur
preprocessor = ResNetDltExtractor(_model, _num_classes)
Note that the number of classes is irrelevant here, but is required to build the
network (before loading it). In this set we just need a dummy `extractor` in Bob pipeline which can be defined in the configuration file in this way:
network (before loading it). ``_model`` and ``_num_classes`` are optional input arguments and will be set automatically. In this set, we just need a dummy ``extractor`` in bob_ pipeline which can be defined in the configuration file in this way:
.. code:: python
......@@ -73,3 +70,6 @@ network (before loading it). In this set we just need a dummy `extractor` in Bob
You can easily implement your own extractor based on your own network too. Just have
a look at the code in ``bob/learn/pytorch/preprocessor/audio``.
.. _bob: http://idiap.github.io/bob/
.. _pytorch: http://pytorch.org/
\ No newline at end of file
......@@ -18,6 +18,7 @@ Users Guide
guide_conv_autoencoder.rst
guide_mlp.rst
guide_mccnn_pad.rst
guide_audio_extractor.rst
================
Reference Manual
......
......@@ -22,3 +22,11 @@ Trainers
.. automodule:: bob.learn.pytorch.trainers
Preprocessors
-------------
.. automodule:: bob.learn.pytorch.preprocessor
Extractors
----------
.. automodule:: bob.learn.pytorch.extractor
\ No newline at end of file
0% Loading or .
You are about to add 0 people to the discussion. Proceed with caution.
Please register or to comment