Update docs (#1550)

* Use `bibtex` for paper citations.
  * add `override.css` for fixing back reference.
  * wav2vec2
  * wav2letter
  * convtasnet
  * deepspeech
  * rnnt-loss
  * griffinlim
* Fix broken references in `filtering`.
* Fix note in soundfile backends.
* Tweak wav2vec2 example.
* Removes unused `pytorch_theme.css`

docs/requirements.txt

@@ -1,4 +1,5 @@
 sphinx==2.4.4
 -e git+git://github.com/pytorch/pytorch_sphinx_theme.git#egg=pytorch_sphinx_theme
 sphinxcontrib.katex
+sphinxcontrib.bibtex
 matplotlib

docs/source/_static/css/override.css

@@ -0,0 +1,8 @@
+/* Fix for bibtex reference */
+dl.footnote.brackets > dt.label > span.brackets > a.fn-backref {
+    position: inherit
+}
+/* Fix for bibtex back reference */
+dl.footnote.brackets > dt.label > span.fn-backref > a {
+    position: inherit
+}

docs/source/conf.py

@@ -41,6 +41,7 @@
     'sphinx.ext.napoleon',
     'sphinx.ext.viewcode',
     'sphinxcontrib.katex',
+    'sphinxcontrib.bibtex',
 ]
 
 # katex options
@@ -55,6 +56,8 @@
 ]
 '''
 
+bibtex_bibfiles = ['refs.bib']
+
 napoleon_use_ivar = True
 napoleon_numpy_docstring = False
 napoleon_google_docstring = True
@@ -133,22 +136,10 @@
 # relative to this directory. They are copied after the builtin static files,
 # so a file named "default.css" will overwrite the builtin "default.css".
 html_static_path = ['_static']
-
-
-def setup(app):
-    # NOTE: in Sphinx 1.8+ `html_css_files` is an official configuration value
-    # and can be moved outside of this function (and the setup(app) function
-    # can be deleted).
-    html_css_files = [
-        'https://cdn.jsdelivr.net/npm/katex@0.10.0-beta/dist/katex.min.css'
-    ]
-
-    # In Sphinx 1.8 it was renamed to `add_css_file`, 1.7 and prior it is
-    # `add_stylesheet` (deprecated in 1.8).
-    add_css = getattr(app, 'add_css_file', app.add_stylesheet)
-    for css_file in html_css_files:
-        add_css(css_file)
-
+html_css_files = [
+    'css/override.css',
+    'https://cdn.jsdelivr.net/npm/katex@0.10.0-beta/dist/katex.min.css'
+]
 
 # -- Options for HTMLHelp output ------------------------------------------
 

docs/source/functional.rst

@@ -235,3 +235,8 @@ vad
 ---------------------------
 
 .. autofunction:: spectral_centroid
+
+References
+~~~~~~~~~~
+
+.. footbibliography::

docs/source/models.rst

@@ -2,43 +2,43 @@
     :class: hidden-section
 
 torchaudio.models
-======================
+=================
 
 .. currentmodule:: torchaudio.models
 
 The models subpackage contains definitions of models for addressing common audio tasks.
 
 
-:hidden:`ConvTasNet`
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+ConvTasNet
+~~~~~~~~~~
 
 .. autoclass:: ConvTasNet
 
   .. automethod:: forward
 
 
-:hidden:`DeepSpeech`
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+DeepSpeech
+~~~~~~~~~~
 
 .. autoclass:: DeepSpeech
 
   .. automethod:: forward
 
 
-:hidden:`Wav2Letter`
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+Wav2Letter
+~~~~~~~~~~
 
 .. autoclass:: Wav2Letter
 
   .. automethod:: forward
 
 
 
-:hidden:`Wav2Vec2.0`
-~~~~~~~~~~~~~~~~~~~~
+Wav2Vec2.0
+~~~~~~~~~~
 
-Model
------
+Wav2Vec2Model
+-------------
 
 .. autoclass:: Wav2Vec2Model
 
@@ -49,26 +49,47 @@ Model
 Factory Functions
 -----------------
 
+wav2vec2_base
+-------------
+
 .. autofunction:: wav2vec2_base
 
+wav2vec2_large
+--------------
+
 .. autofunction:: wav2vec2_large
 
+wav2vec2_large_lv60k
+--------------------
+
 .. autofunction:: wav2vec2_large_lv60k
 
 .. currentmodule:: torchaudio.models.wav2vec2.utils
 
 Utility Functions
 -----------------
 
+import_huggingface_model
+------------------------
+
 .. autofunction:: import_huggingface_model
 
+import_fairseq_model
+--------------------
+
 .. autofunction:: import_fairseq_model
 
 .. currentmodule:: torchaudio.models
 
-:hidden:`WaveRNN`
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+WaveRNN
+~~~~~~~
 
 .. autoclass:: WaveRNN
 
   .. automethod:: forward
+
+References
+~~~~~~~~~~
+
+.. footbibliography::
+

docs/source/refs.bib

@@ -0,0 +1,90 @@
+@misc{baevski2020wav2vec,
+      title={wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations}, 
+      author={Alexei Baevski and Henry Zhou and Abdelrahman Mohamed and Michael Auli},
+      year={2020},
+      eprint={2006.11477},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL}
+}
+@misc{hannun2014deep,
+      title={Deep Speech: Scaling up end-to-end speech recognition}, 
+      author={Awni Hannun and Carl Case and Jared Casper and Bryan Catanzaro and Greg Diamos and Erich Elsen and Ryan Prenger and Sanjeev Satheesh and Shubho Sengupta and Adam Coates and Andrew Y. Ng},
+      year={2014},
+      eprint={1412.5567},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL}
+}
+@misc{graves2012sequence,
+      title={Sequence Transduction with Recurrent Neural Networks}, 
+      author={Alex Graves},
+      year={2012},
+      eprint={1211.3711},
+      archivePrefix={arXiv},
+      primaryClass={cs.NE}
+}
+@misc{collobert2016wav2letter,
+      title={Wav2Letter: an End-to-End ConvNet-based Speech Recognition System}, 
+      author={Ronan Collobert and Christian Puhrsch and Gabriel Synnaeve},
+      year={2016},
+      eprint={1609.03193},
+      archivePrefix={arXiv},
+      primaryClass={cs.LG}
+}
+@misc{kalchbrenner2018efficient,
+      title={Efficient Neural Audio Synthesis}, 
+      author={Nal Kalchbrenner and Erich Elsen and Karen Simonyan and Seb Noury and Norman Casagrande and Edward Lockhart and Florian Stimberg and Aaron van den Oord and Sander Dieleman and Koray Kavukcuoglu},
+      year={2018},
+      eprint={1802.08435},
+      archivePrefix={arXiv},
+      primaryClass={cs.SD}
+}
+@article{Luo_2019,
+   title={Conv-TasNet: Surpassing Ideal Time–Frequency Magnitude Masking for Speech Separation},
+   volume={27},
+   ISSN={2329-9304},
+   url={http://dx.doi.org/10.1109/TASLP.2019.2915167},
+   DOI={10.1109/taslp.2019.2915167},
+   number={8},
+   journal={IEEE/ACM Transactions on Audio, Speech, and Language Processing},
+   publisher={Institute of Electrical and Electronics Engineers (IEEE)},
+   author={Luo, Yi and Mesgarani, Nima},
+   year={2019},
+   month={Aug},
+   pages={1256–1266}
+}
+@InProceedings{ brian_mcfee-proc-scipy-2015,
+  author    = { {B}rian {M}c{F}ee and {C}olin {R}affel and {D}awen {L}iang and {D}aniel {P}.{W}. {E}llis and {M}att {M}c{V}icar and {E}ric {B}attenberg and {O}riol {N}ieto },
+  title     = { librosa: {A}udio and {M}usic {S}ignal {A}nalysis in {P}ython },
+  booktitle = { {P}roceedings of the 14th {P}ython in {S}cience {C}onference },
+  pages     = { 18 - 24 },
+  year      = { 2015 },
+  editor    = { {K}athryn {H}uff and {J}ames {B}ergstra },
+  doi       = { 10.25080/Majora-7b98e3ed-003 }
+}
+@INPROCEEDINGS{6701851,
+  author={Perraudin, Nathanaël and Balazs, Peter and Søndergaard, Peter L.},
+  booktitle={2013 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics}, 
+  title={A fast Griffin-Lim algorithm}, 
+  year={2013},
+  volume={},
+  number={},
+  pages={1-4},
+  doi={10.1109/WASPAA.2013.6701851}}
+@INPROCEEDINGS{1172092,
+  author={Griffin, D. and Jae Lim},
+  booktitle={ICASSP '83. IEEE International Conference on Acoustics, Speech, and Signal Processing}, 
+  title={Signal estimation from modified short-time Fourier transform}, 
+  year={1983},
+  volume={8},
+  number={},
+  pages={804-807},
+  doi={10.1109/ICASSP.1983.1172092}}
+@INPROCEEDINGS{6854049,
+  author={Ghahremani, Pegah and BabaAli, Bagher and Povey, Daniel and Riedhammer, Korbinian and Trmal, Jan and Khudanpur, Sanjeev},
+  booktitle={2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)}, 
+  title={A pitch extraction algorithm tuned for automatic speech recognition}, 
+  year={2014},
+  volume={},
+  number={},
+  pages={2494-2498},
+  doi={10.1109/ICASSP.2014.6854049}}