links, ressources

_doc/sphinxdoc/source/ressources.rst

@@ -49,17 +49,18 @@ Source de données
   Les jeux **UCI** apparaissent dans de nombreux scientifiques.
 * *ML* `Data Science at Microsoft Research <http://research.microsoft.com/en-us/projects/data-science-initiative/default.aspx#datasets>`_
 * *ML* `Ensembles de données publics AWS (Amazon) <https://aws.amazon.com/public-data-sets/>`_, `AWS Public Data Sets <https://aws.amazon.com/datasets/>`_
-* *ML-graphes* `Stanford Large Network Dataset Collection <http://snap.stanford.edu/data/>`_
 * *ML* `Data Analysis, Modeling and Machine Learning Group <http://ama.liglab.fr/resourcestools/datasets/>`_
 * *ML* `Microsoft Research Letor <http://research.microsoft.com/en-us/um/beijing/projects/letor/letor4dataset.aspx>`_
+* *ML* `List of datasets for machine learning research <https://en.wikipedia.org/wiki/List_of_datasets_for_machine_learning_research>`_
+* *ML-deep*: `Open Data for Deep Learning <https://deeplearning4j.org/opendata>`_
+* *ML-graphes* `Stanford Large Network Dataset Collection <http://snap.stanford.edu/data/>`_
+* *ML-big* `Pascal Large Scale Learning Challenge <http://largescale.ml.tu-berlin.de/instructions/>`_
 * *ML-big* `170 millions courses de taxi à New-York <http://chriswhong.com/open-data/foil_nyc_taxi/>`_
   (via l'article `Building Azure ML Models on the NYC Taxi Dataset <http://blogs.technet.com/b/machinelearning/archive/2015/04/02/building-azure-ml-models-on-the-nyc-taxi-dataset.aspx>`_)
 * *ML-text* `urls, spam, ... <http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/binary.html>`_, jeux de données utilisés
   comme benchmark pour la libraire `libsvm <http://www.csie.ntu.edu.tw/~cjlin/libsvm/>`_
-* *ML-big* `Pascal Large Scale Learning Challenge <http://largescale.ml.tu-berlin.de/instructions/>`_
 * *ML-image* `Labeled Faces in the Wild <http://vis-www.cs.umass.edu/lfw/>`_ : 1323 images, 5749 personnes, 1680 personnes avec 2 ou plus d'images,
   lire `How well do facial recognition algorithms cope with a million strangers? <http://www.washington.edu/news/2016/06/23/how-well-do-facial-recognition-algorithms-cope-with-a-million-strangers/>`_
-* *ML* `List of datasets for machine learning research <https://en.wikipedia.org/wiki/List_of_datasets_for_machine_learning_research>`_
 * *musique* `Semantic Artist Similarity Dataset <http://mtg.upf.edu/download/datasets/semantic-similarity>`_
 * *musique* `The Music Matrix – Exploring tags in the Million Song Dataset <http://musicmachinery.com/2011/11/27/the-music-matrix-exploring-tags-in-the-million-song-dataset/>`_
 * *musique* `Audio Content Analysis Datasets <http://www.audiocontentanalysis.org/data-sets/>`_

_doc/sphinxdoc/source/specials/deep_learning.rst

@@ -32,14 +32,16 @@ Algorithmes
 Librairies
 ----------
 
+* `Torch <http://torch.ch/>`_ et surtout `pytorch <http://pytorch.org/>`_
+  dont le design est plus simple que celui des autres.
 * `Caffee <http://caffe.berkeleyvision.org/>`_ (Berkeley)
 * `CNTK <https://www.microsoft.com/en-us/research/product/cognitive-toolkit/>`_ (Microsoft)
 * `deeplearning4j <https://deeplearning4j.org/>`_
 * `mxnet <https://github.com/dmlc/mxnet>`_
 * `PaddlePaddle <https://github.com/PaddlePaddle/Paddle>`_ (Baidu)
 * `TensorFlow <https://www.tensorflow.org/>`_ (Google)
-* `Theano <http://deeplearning.net/software/theano/>`_,
-* `Torch <http://torch.ch/>`_
+* `Theano <http://deeplearning.net/software/theano/>`_ (n'est plus maintenu)
+
 
 `Keras <https://keras.io/>`_ ou `chainer <http://chainer.org/>`_ implémentent des interfaces
 communes pour plusieurs librairies de machine learning.

_doc/sphinxdoc/source/specials/deepproblems.rst

@@ -18,8 +18,9 @@ Deep learning
 GPU
 +++
 
-* `theano <http://deeplearning.net/software/theano/>`_
+* `theano <http://deeplearning.net/software/theano/>`_ (n'est plus maintenu)
 * `cupy <https://github.com/cupy/cupy>`_
+* Tous les modules de deep learning.
 
 Tutoriels
 +++++++++

_doc/sphinxdoc/source/td_2a_mlplus.rst

@@ -610,8 +610,7 @@ Deep Learning
 
 Le deep learning signifie aussi des calculs intensifs et des modules qui
 utilisent un compilateur C++ pour optimiser les calculs
-et le GPU si vous en avez. Il faut lire l'article de blog
-:ref:`Installer theano et un compilateur C++ <blog-install-theno-keras>`.
+et le GPU si vous en avez. 
 Vous pouvez tester votre installation avec le notebook
 :ref:`mldeeppythonrst` ou encore
 `Keras-TensorFlow-GPU-Windows-Installation <https://github.com/antoniosehk/keras-tensorflow-windows-installation>`_.
@@ -726,22 +725,24 @@ Deep Learning en détail
 
 *Modules*
 
-* `theano <http://deeplearning.net/software/theano/>`_
+* `pytorch <http://pytorch.org/>`_ : design plus simple que tous les autres
+* `theano <http://deeplearning.net/software/theano/>`_ (n'est plus maintenu)
+  Il faut lire l'article de blog
+  :ref:`Installer theano et un compilateur C++ <blog-install-theno-keras>`.
 * `keras <https://keras.io/>`_
 * `mxnet <https://github.com/dmlc/mxnet>`_
 * `caffe <http://caffe.berkeleyvision.org/>`_ (`installation <http://caffe.berkeleyvision.org/installation.html>`_)
 * `climin <http://climin.readthedocs.io/en/latest/rmsprop.html>`_ (algorithme de back propagation)
-* `pytorch <http://pytorch.org/>`_ (Facebook)
 * `tensorflow <https://www.tensorflow.org/>`_ (Google)
 * `foolbox <https://github.com/bethgelab/foolbox>`_ :
   trouver des petites perturbations des données qui trompent les réseaux de neurones
+* `cntk <https://github.com/Microsoft/CNTK>`_
 
 *à suivre*
 
 * `chainer <https://github.com/pfnet/chainer>`_
 * `platoon <https://github.com/mila-udem/platoon/>`_ :
-  multi-GPU pour theano
-* `scikit-theano <https://github.com/sklearn-theano/sklearn-theano>`_
+  multi-GPU pour theano (à voir car *theano* n'est plus maintenu)
 * `Federated Learning: Collaborative Machine Learning without Centralized Training Data <https://research.googleblog.com/2017/04/federated-learning-collaborative.html>`_
 
 *Deep learning embarqué*

_doc/sphinxdoc/source/td_3a.rst

@@ -94,7 +94,8 @@ GPU
   `pyopencl <https://documen.tician.de/pyopencl/>`_
   pour ceux qui n'ont pas de carte
   `NVidia <http://www.nvidia.com/content/global/global.php>`_
-* `theano <http://deeplearning.net/software/theano/>`_
+* `theano <http://deeplearning.net/software/theano/>`_ (n'est plus maintenu)
+* Tous les modules de deep learning.
 
 *Bas niveau*