Merge pull request #5718 from notoraptor/prepare-final-release-0.9.0

Prepare final release 0.9.0.

Merge pull request #5718 from notoraptor/prepare-final-release-0.9.0
f3472622 · Frédéric Bastien · GitHub · 898d146d · 8d036e07 · f3472622
--- a/.mailmap
+++ b/.mailmap
@@ -222,3 +222,5 @@ Vivek Kulkarni <viveksck@gmail.com> Vivek Kulkarni <vvkulkarni@cs.stonybrook.edu
 Wei Li <kuantkid@gmail.com> kuantkid <kuantkid@gmail.com>
 Yoshua Bengio <bengioy@iro.umontreal.ca> bengioy@bengio-mac.local <bengioy@bengio-mac.local>
 Ziye Fan <fanziye.cis@gmail.com> FanZiye(t13m) <fanziye.cis@gmail.com>
+Zhouhan LIN <lin.zhouhan@gmail.com> hantek <lin.zhouhan@gmail.com>
+Zhouhan LIN <lin.zhouhan@gmail.com> Zhouhan LIN <hantek@Zhouhans-MacBook-Pro.local>
--- a/HISTORY.txt
+++ b/HISTORY.txt
@@ -5,6 +5,358 @@
 Old Release Notes
 =================

+Theano 0.9.0rc4 (13th of March, 2017)
+=====================================
+
+This release extends the 0.9.0rc3 and announces the upcoming final release 0.9.
+
+Highlights (since 0.9.0rc3):
+ - Documentation updates
+ - DebugMode fixes, cache cleanup fixes and other small fixes
+
+ - New GPU back-end:
+
+   - Fixed offset error in GpuIncSubtensor
+   - Fixed indexing error in GpuAdvancedSubtensor for more than 2 dimensions
+
+A total of 5 people contributed to this release since 0.9.0rc3 and 123 since 0.8.0, see the lists below.
+
+
+Committers since 0.9.0rc3:
+ - Frederic Bastien
+ - Pascal Lamblin
+ - Arnaud Bergeron
+ - Cesar Laurent
+ - Martin Drawitsch
+
+
+Theano 0.9.0rc3 (6th of March, 2017)
+====================================
+
+This release extends the 0.9.0rc2 and announces the upcoming final release 0.9.
+
+Highlights (since 0.9.0rc2):
+ - Graph clean up and faster compilation
+ - New Theano flag conv.assert_shape to check user-provided shapes at runtime (for debugging)
+ - Fix overflow in pooling
+ - Warn if taking softmax over broadcastable dimension
+ - Removed old files not used anymore
+ - Test fixes and crash fixes
+
+ - New GPU back-end:
+
+   - Removed warp-synchronous programming, to get good results with newer CUDA drivers
+
+A total of 5 people contributed to this release since 0.9.0rc2 and 122 since 0.8.0, see the lists below.
+
+
+Committers since 0.9.0rc2:
+ - Frederic Bastien
+ - Arnaud Bergeron
+ - Pascal Lamblin
+ - Florian Bordes
+ - Jan Schlüter
+
+
+Theano 0.9.0rc2 (27th of February, 2017)
+========================================
+
+This release extends the 0.9.0rc1 and announces the upcoming final release 0.9.
+
+Highlights (since 0.9.0rc1):
+ - Fixed dnn conv grad issues
+ - Allowed pooling of empty batch
+ - Use of 64-bit indexing in sparse ops to allow matrix with more then 2\ :sup:`31`\ -1 elements.
+ - Removed old benchmark directory
+ - Crash fixes, bug fixes, warnings improvements, and documentation update
+
+A total of 9 people contributed to this release since 0.9.0rc1 and 121 since 0.8.0, see the lists below.
+
+
+Committers since 0.9.0rc1:
+ - Frederic Bastien
+ - Pascal Lamblin
+ - Steven Bocco
+ - Simon Lefrancois
+ - Lucas Beyer
+ - Michael Harradon
+ - Rebecca N. Palmer
+ - David Bau
+ - Micah Bojrab
+
+
+Theano 0.9.0rc1 (20th of February, 2017)
+========================================
+
+This release extends the 0.9.0beta1 and announces the upcoming final release 0.9.
+
+Highlights (since 0.9.0beta1):
+ - Better integration of Theano+libgpuarray packages into conda distribution
+ - Better handling of Windows end-lines into C codes
+ - Better compatibility with NumPy 1.12
+ - Faster scan optimizations
+ - Fixed broadcast checking in scan
+ - Bug fixes related to merge optimizer and shape inference
+ - many other bug fixes and improvements
+ - Updated documentation
+
+ - New GPU back-end:
+
+   - Value of a shared variable is now set inplace
+
+A total of 26 people contributed to this release since 0.9.0beta1 and 117 since 0.8.0, see the list at the bottom.
+
+Interface changes:
+ - In MRG, replaced method `multinomial_wo_replacement()` with new method `choice()`
+
+Convolution updates:
+ - Implement conv2d_transpose convenience function
+
+GPU:
+ - GPUMultinomialFromUniform op now supports multiple dtypes
+
+New features:
+ - OpFromGraph now allows gradient overriding for every input
+ - Added Abstract Ops for batch normalization that use cuDNN when available and pure Theano CPU/GPU alternatives otherwise
+ - Added new Theano flag cuda.enabled
+ - Added new Theano flag print_global_stats to print some global statistics (time spent) at the end
+
+Others:
+ - Split op now has C code for CPU and GPU
+ - "theano-cache list" now includes compilation times
+
+
+Committers since 0.9.0beta1:
+ - Frederic Bastien
+ - Benjamin Scellier
+ - khaotik
+ - Steven Bocco
+ - Arnaud Bergeron
+ - Pascal Lamblin
+ - Gijs van Tulder
+ - Reyhane Askari
+ - Chinnadhurai Sankar
+ - Vincent Dumoulin
+ - Alexander Matyasko
+ - Cesar Laurent
+ - Nicolas Ballas
+ - affanv14
+ - Faruk Ahmed
+ - Anton Chechetka
+ - Alexandre de Brebisson
+ - Amjad Almahairi
+ - Dimitar Dimitrov
+ - Fuchai
+ - Jan Schlüter
+ - Jonas Degrave
+ - Mathieu Germain
+ - Rebecca N. Palmer
+ - Simon Lefrancois
+ - valtron
+
+
+Theano 0.9.0beta1 (24th of January, 2017)
+=========================================
+
+This release contains a lot of bug fixes and improvements + new features, to prepare the upcoming release candidate.
+
+Highlights:
+ - Many computation and compilation speed up
+ - More numerical stability by default for some graph
+ - Jenkins (gpu tests run on PR in addition to daily buildbot)
+ - Better handling of corner cases for theano functions and graph optimizations
+ - More graph optimization (faster execution and smaller graph, so more readable)
+ - Less c code compilation
+ - Better Python 3.5 support
+ - Better numpy 1.12 support
+ - Support newer Mac and Windows version
+ - Conda packages for Mac, Linux and Windows
+ - Theano scripts now works on Windows
+ - scan with checkpoint (trade off between speed and memory usage, useful for long sequences)
+ - Added a bool dtype
+
+ - New GPU back-end:
+
+   - float16 storage
+   - better mapping between theano device number and nvidia-smi number, using the PCI bus ID of graphic cards
+   - More pooling support on GPU when cuDNN isn't there
+   - ignore_border=False is now implemented for pooling
+
+
+A total of 111 people contributed to this release since 0.8.0, see the list at the bottom.
+
+
+Interface changes:
+ - New pooling interface
+ - Pooling parameters can change at run time
+ - When converting empty list/tuple, now we use floatX dtype
+ - The MRG random generator now try to infer the broadcast pattern of its output
+ - Move softsign out of sandbox to theano.tensor.nnet.softsign
+ - Roll make the shift be modulo the size of the axis we roll on
+ - Merge CumsumOp/CumprodOp into CumOp
+ - round() default to the same as NumPy: half_to_even
+
+Convolution updates:
+ - Multi-cores convolution and pooling on CPU
+ - New abstract 3d convolution interface similar to the 2d convolution interface
+ - Dilated convolution
+
+GPU:
+ - cuDNN: support versoin 5.1 and wrap batch normalization (2d and 3d) and RNN functions
+ - Multiple-GPU, synchrone update (via platoon, use NCCL)
+ - GpuAdvancedSubtensor in new back-end
+ - Gemv(matrix-vector product) speed up for special shape
+ - Support for MaxAndArgMax for some axis combination
+ - Support for solve (using cusolver), erfinv and erfcinv
+ - cublas gemv workaround when we reduce on an axis with a dimensions size of 0
+ - Warn user that some cuDNN algorithms may produce unexpected results in certain environments
+   for convolution backward filter operations
+
+New features:
+ - Add gradient of solve, tensorinv (CPU), tensorsolve (CPU) searchsorted (CPU)
+ - Add Multinomial Without Replacement
+ - conv3d2d support full and half mode (REMOVE?)
+ - Add DownsampleFactorMaxGradGrad.grad
+ - Allow partial evaluation of compiled function
+ - More Rop support
+ - Indexing support ellipsis: a[..., 3], a[1,...,3]
+ - Added theano.tensor.{tensor5,dtensor5, ...}
+ - compiledir_format support device
+ - Added new Theano flag cmodule.age_thresh_use
+
+Others:
+ - Speed up argmax only on gpu (without also needing the max)
+ - A few unfrequent bugfix
+ - More stack trace in error message
+ - Speed up cholesky grad
+ - log(sum(exp(...))) now get stability optimized
+
+Other more detailed changes:
+ - Allow more then one output to be an destructive inplace
+ - Add flag profiling.ignore_first_call, useful to profile the new gpu back-end
+ - Doc/error message fixes/updates
+ - More support of negative axis
+ - Added the keepdims parameter to the norm function
+ - Crash fixes
+ - Make scan gradient more deterministic
+ - Add support for space in path on Windows
+ - remove ProfileMode (use Theano flag profile=True instead)
+
+
+Committers since 0.8.0:
+ - Frederic Bastien
+ - Arnaud Bergeron
+ - Pascal Lamblin
+ - Ramana Subramanyam
+ - Simon Lefrancois
+ - Steven Bocco
+ - Gijs van Tulder
+ - Cesar Laurent
+ - Chiheb Trabelsi
+ - Chinnadhurai Sankar
+ - Mohammad Pezeshki
+ - Reyhane Askari
+ - Alexander Matyasko
+ - Alexandre de Brebisson
+ - Nan Rosemary Ke
+ - Pierre Luc Carrier
+ - Mathieu Germain
+ - Olivier Mastropietro
+ - khaotik
+ - Saizheng Zhang
+ - Thomas George
+ - Iulian Vlad Serban
+ - Benjamin Scellier
+ - Francesco Visin
+ - Caglar
+ - Harm de Vries
+ - Samira Shabanian
+ - Jakub Sygnowski
+ - Samira Ebrahimi Kahou
+ - Mikhail Korobov
+ - Faruk Ahmed
+ - Fei Wang
+ - Jan Schlüter
+ - Kv Manohar
+ - Jesse Livezey
+ - Kelvin Xu
+ - Matt Graham
+ - Ruslana Makovetsky
+ - Sina Honari
+ - Bryn Keller
+ - Ciyong Chen
+ - Nicolas Ballas
+ - Vitaliy Kurlin
+ - Zhouhan LIN
+ - Gokula Krishnan
+ - Kumar Krishna Agrawal
+ - Ozan Çağlayan
+ - Vincent Michalski
+ - Ray Donnelly
+ - Tim Cooijmans
+ - Vincent Dumoulin
+ - happygds
+ - mockingjamie
+ - Amjad Almahairi
+ - Christos Tsirigotis
+ - Ilya Kulikov
+ - RadhikaG
+ - Taesup (TS) Kim
+ - Ying Zhang
+ - Karthik Karanth
+ - Kirill Bobyrev
+ - Yang Zhang
+ - Yaroslav Ganin
+ - Liwei Cai
+ - Morgan Stuart
+ - Tim Gasper
+ - Xavier Bouthillier
+ - p
+ - texot
+ - Andrés Gottlieb
+ - Ben Poole
+ - Bhavishya Pohani
+ - Carl Thomé
+ - Evelyn Mitchell
+ - Fei Zhan
+ - Fábio Perez
+ - Gennadiy Tupitsin
+ - Gilles Louppe
+ - Greg Ciccarelli
+ - He
+ - Huan Zhang
+ - Jonas Degrave
+ - Kaixhin
+ - Kevin Keraudren
+ - Maltimore
+ - Marc-Alexandre Cote
+ - Marco
+ - Marius F. Killinger
+ - Maxim Kochurov
+ - Neil
+ - Nizar Assaf
+ - Rithesh Kumar
+ - Rizky Luthfianto
+ - Robin Millette
+ - Roman Ring
+ - Sander Dieleman
+ - Sebastin Santy
+ - Shawn Tan
+ - Wazeer Zulfikar
+ - Wojciech Głogowski
+ - Yann N. Dauphin
+ - gw0 [http://gw.tnode.com/]
+ - hexahedria
+ - hsintone
+ - jakirkham
+ - joncrall
+ - root
+ - superantichrist
+ - tillahoffmann
+ - wazeerzulfikar
+ - you-n-g
+
+
 Theano 0.8.2 (21th of April, 2016)
 ==================================


--- a/NEWS.txt
+++ b/NEWS.txt
@@ -3,278 +3,164 @@ Release Notes
 =============


-Theano 0.9.0rc4 (13th of March, 2017)
-=====================================
+Theano 0.9.0 (20th of March, 2017)
+==================================

-This release extends the 0.9.0rc3 and announces the upcoming final release 0.9.
+This is a final release of Theano, version ``0.9.0``, with a lot of
+new features, interface changes, improvements and bug fixes.

-Highlights (since 0.9.0rc3):
- - Documentation updates
- - DebugMode fixes, cache cleanup fixes and other small fixes
-
- - New GPU back-end:
-
-   - Fixed offset error in GpuIncSubtensor
-   - Fixed indexing error in GpuAdvancedSubtensor for more than 2 dimensions
-
-A total of 5 people contributed to this release since 0.9.0rc3 and 123 since 0.8.0, see the lists below.
+We recommend that everybody update to this version.

+Highlights (since 0.8.0):
+ - Better Python 3.5 support
+ - Better numpy 1.12 support
+ - Conda packages for Mac, Linux and Windows
+ - Support newer Mac and Windows versions
+ - More Windows integration:

-Committers since 0.9.0rc3:
- - Frederic Bastien
- - Pascal Lamblin
- - Arnaud Bergeron
- - Cesar Laurent
- - Martin Drawitsch
+   - Theano scripts (``theano-cache`` and ``theano-nose``) now works on Windows
+   - Better support for Windows end-lines into C codes
+   - Support for space in paths on Windows

+ - Scan improvements:

-Theano 0.9.0rc3 (6th of March, 2017)
-====================================
+   - More scan optimizations, with faster compilation and gradient computation
+   - Support for checkpoint in scan (trade off between speed and memory usage, useful for long sequences)
+   - Fixed broadcast checking in scan

-This release extends the 0.9.0rc2 and announces the upcoming final release 0.9.
+ - Graphs improvements:

-Highlights (since 0.9.0rc2):
- - Graph clean up and faster compilation
- - New Theano flag conv.assert_shape to check user-provided shapes at runtime (for debugging)
- - Fix overflow in pooling
- - Warn if taking softmax over broadcastable dimension
- - Removed old files not used anymore
- - Test fixes and crash fixes
+   - More numerical stability by default for some graphs
+   - Better handling of corner cases for theano functions and graph optimizations
+   - More graph optimizations with faster compilation and execution
+   - smaller and more readable graph

 - New GPU back-end:

-   - Removed warp-synchronous programming, to get good results with newer CUDA drivers
-
-A total of 5 people contributed to this release since 0.9.0rc2 and 122 since 0.8.0, see the lists below.
-
-
-Committers since 0.9.0rc2:
- - Frederic Bastien
- - Arnaud Bergeron
- - Pascal Lamblin
- - Florian Bordes
- - Jan Schlüter
-
-
-Theano 0.9.0rc2 (27th of February, 2017)
-========================================
-
-This release extends the 0.9.0rc1 and announces the upcoming final release 0.9.
-
-Highlights (since 0.9.0rc1):
- - Fixed dnn conv grad issues
- - Allowed pooling of empty batch
- - Use of 64-bit indexing in sparse ops to allow matrix with more then 2\ :sup:`31`\ -1 elements.
- - Removed old benchmark directory
- - Crash fixes, bug fixes, warnings improvements, and documentation update
-
-A total of 9 people contributed to this release since 0.9.0rc1 and 121 since 0.8.0, see the lists below.
-
-
-Committers since 0.9.0rc1:
- - Frederic Bastien
- - Pascal Lamblin
- - Steven Bocco
- - Simon Lefrancois
- - Lucas Beyer
- - Michael Harradon
- - Rebecca N. Palmer
- - David Bau
- - Micah Bojrab
-
-
-Theano 0.9.0rc1 (20th of February, 2017)
-========================================
-
-This release extends the 0.9.0beta1 and announces the upcoming final release 0.9.
+   - Removed warp-synchronous programming to get good results with newer CUDA drivers
+   - More pooling support on GPU when cuDNN isn't available
+   - Full support of ignore_border option for pooling
+   - Inplace storage for shared variables
+   - float16 storage
+   - Using PCI bus ID of graphic cards for a better mapping between theano device number and nvidia-smi number
+   - Fixed offset error in ``GpuIncSubtensor``

-Highlights (since 0.9.0beta1):
- - Better integration of Theano+libgpuarray packages into conda distribution
- - Better handling of Windows end-lines into C codes
- - Better compatibility with NumPy 1.12
- - Faster scan optimizations
- - Fixed broadcast checking in scan
+ - Less C code compilation
+ - Added support for bool dtype
+ - Updated and more complete documentation
 - Bug fixes related to merge optimizer and shape inference
- - many other bug fixes and improvements
- - Updated documentation
+ - Lot of other bug fixes, crashes fixes and warning improvements

- - New GPU back-end:
-
-   - Value of a shared variable is now set inplace
-
-A total of 26 people contributed to this release since 0.9.0beta1 and 117 since 0.8.0, see the list at the bottom.
+A total of 123 people contributed to this release since 0.8.0, see list below.

 Interface changes:
- - In MRG, replaced method `multinomial_wo_replacement()` with new method `choice()`
-
-Convolution updates:
- - Implement conv2d_transpose convenience function
-
-GPU:
- - GPUMultinomialFromUniform op now supports multiple dtypes
-
-New features:
- - OpFromGraph now allows gradient overriding for every input
- - Added Abstract Ops for batch normalization that use cuDNN when available and pure Theano CPU/GPU alternatives otherwise
- - Added new Theano flag cuda.enabled
- - Added new Theano flag print_global_stats to print some global statistics (time spent) at the end
-
-Others:
- - Split op now has C code for CPU and GPU
- - "theano-cache list" now includes compilation times
+ - Merged ``CumsumOp/CumprodOp`` into ``CumOp``
+ - In MRG module:

+   - Replaced method ``multinomial_wo_replacement()`` with new method ``choice()``
+   - Random generator now tries to infer the broadcast pattern of its output

-Committers since 0.9.0beta1:
- - Frederic Bastien
- - Benjamin Scellier
- - khaotik
- - Steven Bocco
- - Arnaud Bergeron
- - Pascal Lamblin
- - Gijs van Tulder
- - Reyhane Askari
- - Chinnadhurai Sankar
- - Vincent Dumoulin
- - Alexander Matyasko
- - Cesar Laurent
- - Nicolas Ballas
- - affanv14
- - Faruk Ahmed
- - Anton Chechetka
- - Alexandre de Brebisson
- - Amjad Almahairi
- - Dimitar Dimitrov
- - Fuchai
- - Jan Schlüter
- - Jonas Degrave
- - Mathieu Germain
- - Rebecca N. Palmer
- - Simon Lefrancois
- - valtron
-
-
-Theano 0.9.0beta1 (24th of January, 2017)
-=========================================
-
-This release contains a lot of bug fixes and improvements + new features, to prepare the upcoming release candidate.
-
-Highlights:
- - Many computation and compilation speed up
- - More numerical stability by default for some graph
- - Jenkins (gpu tests run on PR in addition to daily buildbot)
- - Better handling of corner cases for theano functions and graph optimizations
- - More graph optimization (faster execution and smaller graph, so more readable)
- - Less c code compilation
- - Better Python 3.5 support
- - Better numpy 1.12 support
- - Support newer Mac and Windows version
- - Conda packages for Mac, Linux and Windows
- - Theano scripts now works on Windows
- - scan with checkpoint (trade off between speed and memory usage, useful for long sequences)
- - Added a bool dtype
-
- - New GPU back-end:
-
-   - float16 storage
-   - better mapping between theano device number and nvidia-smi number, using the PCI bus ID of graphic cards
-   - More pooling support on GPU when cuDNN isn't there
-   - ignore_border=False is now implemented for pooling
-
-
-A total of 111 people contributed to this release since 0.8.0, see the list at the bottom.
-
-
-Interface changes:
 - New pooling interface
 - Pooling parameters can change at run time
- - When converting empty list/tuple, now we use floatX dtype
- - The MRG random generator now try to infer the broadcast pattern of its output
- - Move softsign out of sandbox to theano.tensor.nnet.softsign
- - Roll make the shift be modulo the size of the axis we roll on
- - Merge CumsumOp/CumprodOp into CumOp
- - round() default to the same as NumPy: half_to_even
+ - Moved ``softsign`` out of sandbox to ``theano.tensor.nnet.softsign``
+ - Using floatX dtype when converting empty list/tuple
+ - ``Roll`` make the shift be modulo the size of the axis we roll on
+ - ``round()`` default to the same as NumPy: half_to_even

 Convolution updates:
+ - Support of full and half modes for 2D and 3D convolutions including in ``conv3d2d``
+ - Allowed pooling of empty batch
+ - Implement ``conv2d_transpose`` convenience function
 - Multi-cores convolution and pooling on CPU
 - New abstract 3d convolution interface similar to the 2d convolution interface
 - Dilated convolution

+
 GPU:
 - cuDNN: support versoin 5.1 and wrap batch normalization (2d and 3d) and RNN functions
 - Multiple-GPU, synchrone update (via platoon, use NCCL)
- - GpuAdvancedSubtensor in new back-end
 - Gemv(matrix-vector product) speed up for special shape
- - Support for MaxAndArgMax for some axis combination
- - Support for solve (using cusolver), erfinv and erfcinv
 - cublas gemv workaround when we reduce on an axis with a dimensions size of 0
 - Warn user that some cuDNN algorithms may produce unexpected results in certain environments
   for convolution backward filter operations
+ - ``GPUMultinomialFromUniform`` op now supports multiple dtypes
+ - Support for ``MaxAndArgMax`` for some axis combination
+ - Support for solve (using cusolver), erfinv and erfcinv
+ - Implemented ``GpuAdvancedSubtensor``

 New features:
- - Add gradient of solve, tensorinv (CPU), tensorsolve (CPU) searchsorted (CPU)
- - Add Multinomial Without Replacement
- - conv3d2d support full and half mode (REMOVE?)
- - Add DownsampleFactorMaxGradGrad.grad
- - Allow partial evaluation of compiled function
+ - ``OpFromGraph`` now allows gradient overriding for every input
+ - Added Abstract Ops for batch normalization that use cuDNN when available and pure Theano CPU/GPU alternatives otherwise
+ - Added gradient of solve, tensorinv (CPU), tensorsolve (CPU), searchsorted (CPU), DownsampleFactorMaxGradGrad (CPU)
+ - Added Multinomial Without Replacement
+ - Allowed partial evaluation of compiled function
 - More Rop support
- - Indexing support ellipsis: a[..., 3], a[1,...,3]
- - Added theano.tensor.{tensor5,dtensor5, ...}
+ - Indexing support ellipsis: ``a[..., 3]```, ``a[1,...,3]``
+ - Added ``theano.tensor.{tensor5,dtensor5, ...}``
 - compiledir_format support device
- - Added new Theano flag cmodule.age_thresh_use
+ - Added New Theano flag ``conv.assert_shape`` to check user-provided shapes at runtime (for debugging)
+ - Added new Theano flag ``cmodule.age_thresh_use``
+ - Added new Theano flag ``cuda.enabled``
+ - Added new Theano flag ``nvcc.cudafe`` to enable faster compilation and import with old CUDA back-end
+ - Added new Theano flag ``print_global_stats`` to print some global statistics (time spent) at the end
+ - Added new Theano flag ``profiling.ignore_first_call``, useful to profile the new gpu back-end
+ - remove ProfileMode (use Theano flag ``profile=True`` instead)
+

 Others:
- - Speed up argmax only on gpu (without also needing the max)
- - A few unfrequent bugfix
- - More stack trace in error message
+ - Split op now has C code for CPU and GPU
+ - ``theano-cache list`` now includes compilation times
+ - Speed up argmax only on GPU (without also needing the max)
+ - More stack trace in error messages
 - Speed up cholesky grad
- - log(sum(exp(...))) now get stability optimized
+ - ``log(sum(exp(...)))`` now get stability optimized
+

 Other more detailed changes:
- - Allow more then one output to be an destructive inplace
- - Add flag profiling.ignore_first_call, useful to profile the new gpu back-end
- - Doc/error message fixes/updates
+ - Added Jenkins (gpu tests run on pull requests in addition to daily buildbot)
+ - Removed old benchmark directory and other old files not used anymore
+ - Use of 64-bit indexing in sparse ops to allow matrix with more then 2\ :sup:`31`\ -1 elements
+ - Allowed more then one output to be an destructive inplace
 - More support of negative axis
 - Added the keepdims parameter to the norm function
- - Crash fixes
 - Make scan gradient more deterministic
- - Add support for space in path on Windows
- - remove ProfileMode (use Theano flag profile=True instead)

-
-Committers since 0.8.0:
+Commiters since 0.8.0:
 - Frederic Bastien
 - Arnaud Bergeron
 - Pascal Lamblin
+ - Steven Bocco
 - Ramana Subramanyam
 - Simon Lefrancois
- - Steven Bocco
 - Gijs van Tulder
- - Cesar Laurent
+ - Benjamin Scellier
+ - khaotik
 - Chiheb Trabelsi
 - Chinnadhurai Sankar
- - Mohammad Pezeshki
+ - Cesar Laurent
 - Reyhane Askari
+ - Mohammad Pezeshki
 - Alexander Matyasko
 - Alexandre de Brebisson
+ - Mathieu Germain
 - Nan Rosemary Ke
 - Pierre Luc Carrier
- - Mathieu Germain
 - Olivier Mastropietro
- - khaotik
- - Saizheng Zhang
 - Thomas George
+ - Saizheng Zhang
 - Iulian Vlad Serban
- - Benjamin Scellier
 - Francesco Visin
 - Caglar
+ - Faruk Ahmed
 - Harm de Vries
 - Samira Shabanian
+ - Vincent Dumoulin
+ - Nicolas Ballas
 - Jakub Sygnowski
+ - Jan Schlüter
 - Samira Ebrahimi Kahou
 - Mikhail Korobov
- - Faruk Ahmed
 - Fei Wang
- - Jan Schlüter
 - Kv Manohar
 - Jesse Livezey
 - Kelvin Xu
@@ -283,29 +169,34 @@ Committers since 0.8.0:
 - Sina Honari
 - Bryn Keller
 - Ciyong Chen
- - Nicolas Ballas
 - Vitaliy Kurlin
 - Zhouhan LIN
 - Gokula Krishnan
 - Kumar Krishna Agrawal
 - Ozan Çağlayan
 - Vincent Michalski
+ - affanv14
+ - Amjad Almahairi
 - Ray Donnelly
 - Tim Cooijmans
- - Vincent Dumoulin
 - happygds
 - mockingjamie
- - Amjad Almahairi
 - Christos Tsirigotis
+ - Florian Bordes
 - Ilya Kulikov
 - RadhikaG
 - Taesup (TS) Kim
 - Ying Zhang
+ - Anton Chechetka
 - Karthik Karanth
 - Kirill Bobyrev
+ - Rebecca N. Palmer
 - Yang Zhang
 - Yaroslav Ganin
+ - Jonas Degrave
 - Liwei Cai
+ - Lucas Beyer
+ - Michael Harradon
 - Morgan Stuart
 - Tim Gasper
 - Xavier Bouthillier
@@ -315,22 +206,26 @@ Committers since 0.8.0:
 - Ben Poole
 - Bhavishya Pohani
 - Carl Thomé
+ - David Bau
+ - Dimitar Dimitrov
 - Evelyn Mitchell
 - Fei Zhan
+ - Fuchai
 - Fábio Perez
 - Gennadiy Tupitsin
 - Gilles Louppe
 - Greg Ciccarelli
 - He
 - Huan Zhang
- - Jonas Degrave
 - Kaixhin
 - Kevin Keraudren
 - Maltimore
 - Marc-Alexandre Cote
 - Marco
 - Marius F. Killinger
+ - Martin Drawitsch
 - Maxim Kochurov
+ - Micah Bojrab
 - Neil
 - Nizar Assaf
 - Rithesh Kumar
@@ -351,5 +246,6 @@ Committers since 0.8.0:
 - root
 - superantichrist
 - tillahoffmann
+ - valtron
 - wazeerzulfikar
 - you-n-g
--- a/NEWS_DEV.txt
+++ b/NEWS_DEV.txt
@@ -4,10 +4,10 @@
 DRAFT Release Notes
 ===================

-git log -p rel-0.8.0... |grep Merge|grep '#[0123456789]' |cut -f 8 -d ' ' | sed 's\#\* https://github.com/Theano/Theano/pull/\'
+git log -p rel-0.9.0... |grep Merge|grep '#[0123456789]' |cut -f 8 -d ' ' | sed 's\#\* https://github.com/Theano/Theano/pull/\'

 # Commit count per user
-git shortlog -sn rel-0.8.0..
+git shortlog -sn rel-0.9.0..



@@ -15,670 +15,62 @@ git shortlog -sn rel-0.8.0..

 TODO: better Theano conv doc

+# NB: Following notes contains infos since 0.9.0.
+
 Highlights:
- - Better integration of Theano+libgpuarray packages into conda distribution
- - Better handling of Windows end-lines into C codes
- - Better compatibility with NumPy 1.12
- - Faster scan optimizations
- - Fixed broadcast checking in scan
- - Bug fixes related to merge optimizer and shape inference
- - many other bug fixes and improvements
- - Updated documentation
- - Many computation and compilation speed up
- - More numerical stability by default for some graph
- - Jenkins (gpu tests run on PR in addition to daily buildbot)
- - Better handling of corner cases for theano functions and graph optimizations
- - More graph optimization (faster execution and smaller graph, so more readable)
- - Less c code compilation
- - Better Python 3.5 support
- - Better numpy 1.12 support
- - Support newer Mac and Windows version
- - Conda packages for Mac, Linux and Windows
- - Theano scripts now works on Windows
- - scan with checkpoint (trade off between speed and memory usage, useful for long sequences)
- - Added a bool dtype
+ - Speed up graph toposort algorithm
+ - Speed up Theano scan compilation and gradient computation
+ - Added meaningful message when missing inputs to scan
+ - Bug fixes related to Debug mode

 - New GPU back-end:

-   - Fixed offset error in GpuIncSubtensor
-   - Fixed indexing error in GpuAdvancedSubtensor for more than 2 dimensions
-   - Value of a shared variable is now set inplace
-   - float16 storage
-   - better mapping between theano device number and nvidia-smi number, using the PCI bus ID of graphic cards
-   - More pooling support on GPU when cuDNN isn't there
-   - ignore_border=False is now implemented for pooling
-   - Removed warp-synchronous programming
+   - Added useful stats for GPU in profile mode
+   - Added documentation for GPU float16 ops
+

 Interface changes:
- - In MRG, replaced method `multinomial_wo_replacement()` with new method `choice()`
- - New pooling interface
- - Pooling parameters can change at run time
- - When converting empty list/tuple, now we use floatX dtype
- - The MRG random generator now try to infer the broadcast pattern of its output
- - Move softsign out of sandbox to theano.tensor.nnet.softsign
- - Roll make the shift be modulo the size of the axis we roll on
- - Merge CumsumOp/CumprodOp into CumOp
- - round() default to the same as NumPy: half_to_even
+ - Changed ``grad()`` method to ``L_op`` in ops that need the outputs to compute gradient
+ - Merged duplicated diagonal functions into two ops: ``ExtractDiag`` (extract a diagonal to a vector),
+   and ``AllocDiag`` (set a vector as a diagonal of an empty array)

 Convolution updates:
- - Allowed pooling of empty batch
- - Implement conv2d_transpose convenience function
- - Multi-cores convolution and pooling on CPU
- - New abstract 3d convolution interface similar to the 2d convolution interface
- - Dilated convolution
+ - Extended Theano flag ``dnn.enabled`` with new option ``no_check`` to help speed up cuDNN importation


 GPU:
- - GPUMultinomialFromUniform op now supports multiple dtypes
- - cuDNN: support versoin 5.1 and wrap batch normalization (2d and 3d) and RNN functions
- - Multiple-GPU, synchrone update (via platoon, use NCCL)
- - GpuAdvancedSubtensor in new back-end
- - Gemv(matrix-vector product) speed up for special shape
- - Support for MaxAndArgMax for some axis combination
- - Support for solve (using cusolver), erfinv and erfcinv
- - cublas gemv workaround when we reduce on an axis with a dimensions size of 0
- - Warn user that some cuDNN algorithms may produce unexpected results in certain environments
-   for convolution backward filter operations
+ - ...

 New features:
- - Added new Theano flag conv.assert_shape
- - OpFromGraph now allows gradient overriding for every input
- - Added Abstract Ops for batch normalization that use cuDNN when available and pure Theano CPU/GPU alternatives otherwise
- - Added new Theano flag cuda.enabled
- - Added new Theano flag print_global_stats to print some global statistics (time spent) at the end
- - Add gradient of solve, tensorinv (CPU), tensorsolve (CPU) searchsorted (CPU)
- - Add Multinomial Without Replacement
- - conv3d2d support full and half mode (REMOVE?)
- - Add DownsampleFactorMaxGradGrad.grad
- - Allow partial evaluation of compiled function
- - More Rop support
- - Indexing support ellipsis: a[..., 3], a[1,...,3]
- - Added theano.tensor.{tensor5,dtensor5, ...}
- - compiledir_format support device
- - Added new Theano flag cmodule.age_thresh_use
+ - Added scalar and elemwise ops for modified Bessel function of order 0 and 1 from ``scipy.special``


 Others:
- - Split op now has C code for CPU and GPU
- - "theano-cache list" now includes compilation times
- - Speed up argmax only on gpu (without also needing the max)
- - A few unfrequent bugfix 
- - More stack trace in error message
- - Speed up cholesky grad
- - log(sum(exp(...))) now get stability optimized
+ - ...


 Other more detailed changes:
- - Added new Theano flag nvcc.cudafe to enable faster compilation and import with old CUDA back-end
- - Use of 64-bit indexing in sparse ops to allow matrix with more then 2\ :sup:`31`\ -1 elements.
- - Allow more then one output to be an destructive inplace
- - Add flag profiling.ignore_first_call, useful to profile the new gpu back-end
- - Doc/error message fixes/updates
- - More support of negative axis
- - Added the keepdims parameter to the norm function
- - Crash fixes
- - Make scan gradient more deterministic
- - Add support for space in path on Windows
- - remove ProfileMode (use Theano flag profile=True instead)
+ - Removed useless warning when profile is manually disabled
+ - Added tests for abstract conv



 ALL THE PR BELLOW HAVE BEEN CHECKED
-* https://github.com/Theano/Theano/pull/5693
-* https://github.com/Theano/Theano/pull/5702
-* https://github.com/Theano/Theano/pull/5697
-* https://github.com/Theano/Theano/pull/5684
-* https://github.com/Theano/Theano/pull/5673
-* https://github.com/Theano/Theano/pull/5680
-* https://github.com/Theano/Theano/pull/5678
-* https://github.com/Theano/Theano/pull/5675
-* https://github.com/Theano/Theano/pull/5666
-* https://github.com/Theano/Theano/pull/5643
-* https://github.com/Theano/Theano/pull/5631
-* https://github.com/Theano/Theano/pull/5651
-* https://github.com/Theano/Theano/pull/5652
-* https://github.com/Theano/Theano/pull/5646
-* https://github.com/Theano/Theano/pull/5634
-* https://github.com/Theano/Theano/pull/5650
-* https://github.com/Theano/Theano/pull/5636
-* https://github.com/Theano/Theano/pull/5635
-* https://github.com/Theano/Theano/pull/5632
-* https://github.com/Theano/Theano/pull/5626
-* https://github.com/Theano/Theano/pull/5625
-* https://github.com/Theano/Theano/pull/5616
-* https://github.com/Theano/Theano/pull/5619
-* https://github.com/Theano/Theano/pull/5614
-* https://github.com/Theano/Theano/pull/5605
-* https://github.com/Theano/Theano/pull/5599
-* https://github.com/Theano/Theano/pull/5609
-* https://github.com/Theano/Theano/pull/5597
-* https://github.com/Theano/Theano/pull/5596
-* https://github.com/Theano/Theano/pull/5550
-* https://github.com/Theano/Theano/pull/5595
-* https://github.com/Theano/Theano/pull/5537
-* https://github.com/Theano/Theano/pull/5593
-* https://github.com/Theano/Theano/pull/5576
-* https://github.com/Theano/Theano/pull/5577
-* https://github.com/Theano/Theano/pull/5587
-* https://github.com/Theano/Theano/pull/5559
-* https://github.com/Theano/Theano/pull/5568
-* https://github.com/Theano/Theano/pull/5553
-* https://github.com/Theano/Theano/pull/5561
-* https://github.com/Theano/Theano/pull/5558
-* https://github.com/Theano/Theano/pull/5544
-* https://github.com/Theano/Theano/pull/5552
-* https://github.com/Theano/Theano/pull/5547
-* https://github.com/Theano/Theano/pull/5542
-* https://github.com/Theano/Theano/pull/5451
-* https://github.com/Theano/Theano/pull/5520
-* https://github.com/Theano/Theano/pull/5539
-* https://github.com/Theano/Theano/pull/5532
-* https://github.com/Theano/Theano/pull/5457
-* https://github.com/Theano/Theano/pull/5477
-* https://github.com/Theano/Theano/pull/5499
-* https://github.com/Theano/Theano/pull/5518
-* https://github.com/Theano/Theano/pull/5527
-* https://github.com/Theano/Theano/pull/5522
-* https://github.com/Theano/Theano/pull/5505
-* https://github.com/Theano/Theano/pull/5523
-* https://github.com/Theano/Theano/pull/5516
-* https://github.com/Theano/Theano/pull/5511
-* https://github.com/Theano/Theano/pull/5443
-* https://github.com/Theano/Theano/pull/5255
-* https://github.com/Theano/Theano/pull/5508
-* https://github.com/Theano/Theano/pull/5479
-* https://github.com/Theano/Theano/pull/5462
-* https://github.com/Theano/Theano/pull/5490
-* https://github.com/Theano/Theano/pull/5480
-* https://github.com/Theano/Theano/pull/5497
-* https://github.com/Theano/Theano/pull/5489
-* https://github.com/Theano/Theano/pull/5474
-* https://github.com/Theano/Theano/pull/5454
-* https://github.com/Theano/Theano/pull/5469
-* https://github.com/Theano/Theano/pull/5458
-* https://github.com/Theano/Theano/pull/5481
-* https://github.com/Theano/Theano/pull/5190
-* https://github.com/Theano/Theano/pull/5473
-* https://github.com/Theano/Theano/pull/5456
-* https://github.com/Theano/Theano/pull/5398
-* https://github.com/Theano/Theano/pull/5468
-* https://github.com/Theano/Theano/pull/5459
-* https://github.com/Theano/Theano/pull/5452
-* https://github.com/Theano/Theano/pull/5298
-* https://github.com/Theano/Theano/pull/5442
-* https://github.com/Theano/Theano/pull/5450
-* https://github.com/Theano/Theano/pull/5435
-* https://github.com/Theano/Theano/pull/5446
-* https://github.com/Theano/Theano/pull/5447
-* https://github.com/Theano/Theano/pull/5445
-* https://github.com/Theano/Theano/pull/5323
-* https://github.com/Theano/Theano/pull/5421
-* https://github.com/Theano/Theano/pull/5343
-* https://github.com/Theano/Theano/pull/5437
-* https://github.com/Theano/Theano/pull/5441
-* https://github.com/Theano/Theano/pull/4659
-* https://github.com/Theano/Theano/pull/5419
-* https://github.com/Theano/Theano/pull/5436
-* https://github.com/Theano/Theano/pull/5427
-* https://github.com/Theano/Theano/pull/5434
-* https://github.com/Theano/Theano/pull/5428
-* https://github.com/Theano/Theano/pull/5412
-* https://github.com/Theano/Theano/pull/5417
-* https://github.com/Theano/Theano/pull/5420
-* https://github.com/Theano/Theano/pull/5422
-* https://github.com/Theano/Theano/pull/5406
-* https://github.com/Theano/Theano/pull/5403
-* https://github.com/Theano/Theano/pull/5410
-* https://github.com/Theano/Theano/pull/5415
-* https://github.com/Theano/Theano/pull/5407
-* https://github.com/Theano/Theano/pull/5321
-* https://github.com/Theano/Theano/pull/5405
-* https://github.com/Theano/Theano/pull/5400
-* https://github.com/Theano/Theano/pull/5346
-* https://github.com/Theano/Theano/pull/5380
-* https://github.com/Theano/Theano/pull/5399
-* https://github.com/Theano/Theano/pull/5182
-* https://github.com/Theano/Theano/pull/5389
-* https://github.com/Theano/Theano/pull/5351
-* https://github.com/Theano/Theano/pull/5377
-* https://github.com/Theano/Theano/pull/5385
-* https://github.com/Theano/Theano/pull/5390
-* https://github.com/Theano/Theano/pull/5387
-* https://github.com/Theano/Theano/pull/5373
-* https://github.com/Theano/Theano/pull/5357
-* https://github.com/Theano/Theano/pull/5363
-* https://github.com/Theano/Theano/pull/5320
-* https://github.com/Theano/Theano/pull/5359
-* https://github.com/Theano/Theano/pull/5362
-* https://github.com/Theano/Theano/pull/5358
-* https://github.com/Theano/Theano/pull/5350
-* https://github.com/Theano/Theano/pull/5317
-* https://github.com/Theano/Theano/pull/5356
-* https://github.com/Theano/Theano/pull/5355
-* https://github.com/Theano/Theano/pull/5353
-* https://github.com/Theano/Theano/pull/5352
-* https://github.com/Theano/Theano/pull/5349
-* https://github.com/Theano/Theano/pull/5267
-* https://github.com/Theano/Theano/pull/5342
-* https://github.com/Theano/Theano/pull/5336
-* https://github.com/Theano/Theano/pull/5335
-* https://github.com/Theano/Theano/pull/5301
-* https://github.com/Theano/Theano/pull/5332
-* https://github.com/Theano/Theano/pull/5299
-* https://github.com/Theano/Theano/pull/5331
-* https://github.com/Theano/Theano/pull/5314
-* https://github.com/Theano/Theano/pull/5276
-* https://github.com/Theano/Theano/pull/5318
-* https://github.com/Theano/Theano/pull/5315
-* https://github.com/Theano/Theano/pull/5313
-* https://github.com/Theano/Theano/pull/5311
-* https://github.com/Theano/Theano/pull/5309
-* https://github.com/Theano/Theano/pull/5306
-* https://github.com/Theano/Theano/pull/5289
-* https://github.com/Theano/Theano/pull/5307
-* https://github.com/Theano/Theano/pull/5290
-* https://github.com/Theano/Theano/pull/5303
-* https://github.com/Theano/Theano/pull/5302
-* https://github.com/Theano/Theano/pull/5300
-* https://github.com/Theano/Theano/pull/5304
-* https://github.com/Theano/Theano/pull/5271
-* https://github.com/Theano/Theano/pull/5247
-* https://github.com/Theano/Theano/pull/5288
-* https://github.com/Theano/Theano/pull/5265
-* https://github.com/Theano/Theano/pull/5223
-* https://github.com/Theano/Theano/pull/5297
-* https://github.com/Theano/Theano/pull/5293
-* https://github.com/Theano/Theano/pull/5286
-* https://github.com/Theano/Theano/pull/5287
-* https://github.com/Theano/Theano/pull/5198
-* https://github.com/Theano/Theano/pull/5282
-* https://github.com/Theano/Theano/pull/5254
-* https://github.com/Theano/Theano/pull/5281
-* https://github.com/Theano/Theano/pull/5231
-* https://github.com/Theano/Theano/pull/5239
-* https://github.com/Theano/Theano/pull/5266
-* https://github.com/Theano/Theano/pull/5277
-* https://github.com/Theano/Theano/pull/5275
-* https://github.com/Theano/Theano/pull/5153
-* https://github.com/Theano/Theano/pull/5269
-* https://github.com/Theano/Theano/pull/5270
-* https://github.com/Theano/Theano/pull/5257
-* https://github.com/Theano/Theano/pull/5245
-* https://github.com/Theano/Theano/pull/5244
-* https://github.com/Theano/Theano/pull/5241
-* https://github.com/Theano/Theano/pull/5251
-* https://github.com/Theano/Theano/pull/5250
-* https://github.com/Theano/Theano/pull/5234
-* https://github.com/Theano/Theano/pull/5242
-* https://github.com/Theano/Theano/pull/5210
-* https://github.com/Theano/Theano/pull/5240
-* https://github.com/Theano/Theano/pull/5229
-* https://github.com/Theano/Theano/pull/5237
-* https://github.com/Theano/Theano/pull/5226
-* https://github.com/Theano/Theano/pull/5212
-* https://github.com/Theano/Theano/pull/4749
-* https://github.com/Theano/Theano/pull/5235
-* https://github.com/Theano/Theano/pull/5214
-* https://github.com/Theano/Theano/pull/5232
-* https://github.com/Theano/Theano/pull/5211
-* https://github.com/Theano/Theano/pull/5205
-* https://github.com/Theano/Theano/pull/5168
-* https://github.com/Theano/Theano/pull/5208
-* https://github.com/Theano/Theano/pull/5199
-* https://github.com/Theano/Theano/pull/5174
-* https://github.com/Theano/Theano/pull/5196
-* https://github.com/Theano/Theano/pull/5185
-* https://github.com/Theano/Theano/pull/5194
-* https://github.com/Theano/Theano/pull/5155
-* https://github.com/Theano/Theano/pull/5193
-* https://github.com/Theano/Theano/pull/5152
-* https://github.com/Theano/Theano/pull/5175
-* https://github.com/Theano/Theano/pull/4355
-* https://github.com/Theano/Theano/pull/5179
-* https://github.com/Theano/Theano/pull/5181
-* https://github.com/Theano/Theano/pull/5132
-* https://github.com/Theano/Theano/pull/5178
-* https://github.com/Theano/Theano/pull/5077
-* https://github.com/Theano/Theano/pull/5159
-* https://github.com/Theano/Theano/pull/5173
-* https://github.com/Theano/Theano/pull/5172
-* https://github.com/Theano/Theano/pull/5164
-* https://github.com/Theano/Theano/pull/5163
-* https://github.com/Theano/Theano/pull/5165
-* https://github.com/Theano/Theano/pull/5166
-* https://github.com/Theano/Theano/pull/5097
-* https://github.com/Theano/Theano/pull/5142
-* https://github.com/Theano/Theano/pull/5133
-* https://github.com/Theano/Theano/pull/5076
-* https://github.com/Theano/Theano/pull/5091
-* https://github.com/Theano/Theano/pull/5160
-* https://github.com/Theano/Theano/pull/5144
-* https://github.com/Theano/Theano/pull/5146
-* https://github.com/Theano/Theano/pull/5119
-* https://github.com/Theano/Theano/pull/5134
-* https://github.com/Theano/Theano/pull/5098
-* https://github.com/Theano/Theano/pull/4736
-* https://github.com/Theano/Theano/pull/5126
-* https://github.com/Theano/Theano/pull/5122
-* https://github.com/Theano/Theano/pull/5086
-* https://github.com/Theano/Theano/pull/4996
-* https://github.com/Theano/Theano/pull/5108
-* https://github.com/Theano/Theano/pull/5105
-* https://github.com/Theano/Theano/pull/4915
-* https://github.com/Theano/Theano/pull/5103
-* https://github.com/Theano/Theano/pull/5096
-* https://github.com/Theano/Theano/pull/5069
-* https://github.com/Theano/Theano/pull/5083
-* https://github.com/Theano/Theano/pull/5073
-* https://github.com/Theano/Theano/pull/5082
-* https://github.com/Theano/Theano/pull/5093
-* https://github.com/Theano/Theano/pull/5084
-* https://github.com/Theano/Theano/pull/5089
-* https://github.com/Theano/Theano/pull/5090
-* https://github.com/Theano/Theano/pull/4948
-* https://github.com/Theano/Theano/pull/5087
-* https://github.com/Theano/Theano/pull/4884
-* https://github.com/Theano/Theano/pull/4862
-* https://github.com/Theano/Theano/pull/5070
-* https://github.com/Theano/Theano/pull/5054
-* https://github.com/Theano/Theano/pull/5060
-* https://github.com/Theano/Theano/pull/5065
-* https://github.com/Theano/Theano/pull/5049
-* https://github.com/Theano/Theano/pull/5062
-* https://github.com/Theano/Theano/pull/5061
-* https://github.com/Theano/Theano/pull/5058
-* https://github.com/Theano/Theano/pull/5057
-* https://github.com/Theano/Theano/pull/4763
-* https://github.com/Theano/Theano/pull/4961
-* https://github.com/Theano/Theano/pull/5050
-* https://github.com/Theano/Theano/pull/5045
-* https://github.com/Theano/Theano/pull/4039
-* https://github.com/Theano/Theano/pull/5028
-* https://github.com/Theano/Theano/pull/4935
-* https://github.com/Theano/Theano/pull/5033
-* https://github.com/Theano/Theano/pull/5038
-* https://github.com/Theano/Theano/pull/5032
-* https://github.com/Theano/Theano/pull/5037
-* https://github.com/Theano/Theano/pull/5031
-* https://github.com/Theano/Theano/pull/5029
-* https://github.com/Theano/Theano/pull/5010
-* https://github.com/Theano/Theano/pull/5014
-* https://github.com/Theano/Theano/pull/5025
-* https://github.com/Theano/Theano/pull/5027
-* https://github.com/Theano/Theano/pull/4992
-* https://github.com/Theano/Theano/pull/5013
-* https://github.com/Theano/Theano/pull/5015
-* https://github.com/Theano/Theano/pull/5023
-* https://github.com/Theano/Theano/pull/4850
-* https://github.com/Theano/Theano/pull/5001
-* https://github.com/Theano/Theano/pull/5002
-* https://github.com/Theano/Theano/pull/4984
-* https://github.com/Theano/Theano/pull/4997
-* https://github.com/Theano/Theano/pull/4991
-* https://github.com/Theano/Theano/pull/4995
-* https://github.com/Theano/Theano/pull/4985
-* https://github.com/Theano/Theano/pull/4972
-* https://github.com/Theano/Theano/pull/4986
-* https://github.com/Theano/Theano/pull/4976
-* https://github.com/Theano/Theano/pull/4951
-* https://github.com/Theano/Theano/pull/4983
-* https://github.com/Theano/Theano/pull/4977
-* https://github.com/Theano/Theano/pull/4968
-* https://github.com/Theano/Theano/pull/4924
-* https://github.com/Theano/Theano/pull/4971
-* https://github.com/Theano/Theano/pull/4876
-* https://github.com/Theano/Theano/pull/4967
-* https://github.com/Theano/Theano/pull/4966
-* https://github.com/Theano/Theano/pull/4964
-* https://github.com/Theano/Theano/pull/4965
-* https://github.com/Theano/Theano/pull/4932
-* https://github.com/Theano/Theano/pull/4958
-* https://github.com/Theano/Theano/pull/4953
-* https://github.com/Theano/Theano/pull/4957
-* https://github.com/Theano/Theano/pull/4956
-* https://github.com/Theano/Theano/pull/4954
-* https://github.com/Theano/Theano/pull/4949
-* https://github.com/Theano/Theano/pull/4950
-* https://github.com/Theano/Theano/pull/4937
-* https://github.com/Theano/Theano/pull/4911
-* https://github.com/Theano/Theano/pull/4947
-* https://github.com/Theano/Theano/pull/4944
-* https://github.com/Theano/Theano/pull/4925
-* https://github.com/Theano/Theano/pull/4941
-* https://github.com/Theano/Theano/pull/4934
-* https://github.com/Theano/Theano/pull/4750
-* https://github.com/Theano/Theano/pull/4756
-* https://github.com/Theano/Theano/pull/4930
-* https://github.com/Theano/Theano/pull/4923
-* https://github.com/Theano/Theano/pull/4920
-* https://github.com/Theano/Theano/pull/4529
-* https://github.com/Theano/Theano/pull/4906
-* https://github.com/Theano/Theano/pull/4913
-* https://github.com/Theano/Theano/pull/4894
-* https://github.com/Theano/Theano/pull/4898
-* https://github.com/Theano/Theano/pull/4773
-* https://github.com/Theano/Theano/pull/4886
-* https://github.com/Theano/Theano/pull/4681
-* https://github.com/Theano/Theano/pull/4888
-* https://github.com/Theano/Theano/pull/4901
-* https://github.com/Theano/Theano/pull/4814
-* https://github.com/Theano/Theano/pull/4810
-* https://github.com/Theano/Theano/pull/4893
-* https://github.com/Theano/Theano/pull/4891
-* https://github.com/Theano/Theano/pull/4772
-* https://github.com/Theano/Theano/pull/4883
-* https://github.com/Theano/Theano/pull/4874
-* https://github.com/Theano/Theano/pull/4852
-* https://github.com/Theano/Theano/pull/4881
-* https://github.com/Theano/Theano/pull/4873
-* https://github.com/Theano/Theano/pull/4871
-* https://github.com/Theano/Theano/pull/4872
-* https://github.com/Theano/Theano/pull/4870
-* https://github.com/Theano/Theano/pull/4865
-* https://github.com/Theano/Theano/pull/4843
-* https://github.com/Theano/Theano/pull/4831
-* https://github.com/Theano/Theano/pull/4867
-* https://github.com/Theano/Theano/pull/4846
-* https://github.com/Theano/Theano/pull/4866
-* https://github.com/Theano/Theano/pull/4864
-* https://github.com/Theano/Theano/pull/4861
-* https://github.com/Theano/Theano/pull/4853
-* https://github.com/Theano/Theano/pull/4856
-* https://github.com/Theano/Theano/pull/4855
-* https://github.com/Theano/Theano/pull/4848
-* https://github.com/Theano/Theano/pull/4824
-* https://github.com/Theano/Theano/pull/4835
-* https://github.com/Theano/Theano/pull/4838
-* https://github.com/Theano/Theano/pull/4841
-* https://github.com/Theano/Theano/pull/4842
-* https://github.com/Theano/Theano/pull/4826
-* https://github.com/Theano/Theano/pull/4798
-* https://github.com/Theano/Theano/pull/4830
-* https://github.com/Theano/Theano/pull/4821
-* https://github.com/Theano/Theano/pull/4808
-* https://github.com/Theano/Theano/pull/4822
-* https://github.com/Theano/Theano/pull/4820
-* https://github.com/Theano/Theano/pull/4804
-* https://github.com/Theano/Theano/pull/4815
-* https://github.com/Theano/Theano/pull/4806
-* https://github.com/Theano/Theano/pull/4802
-* https://github.com/Theano/Theano/pull/4788
-* https://github.com/Theano/Theano/pull/4799
-* https://github.com/Theano/Theano/pull/4669
-* https://github.com/Theano/Theano/pull/4768
-* https://github.com/Theano/Theano/pull/4769
-* https://github.com/Theano/Theano/pull/4771
-* https://github.com/Theano/Theano/pull/4764
-* https://github.com/Theano/Theano/pull/4757
-* https://github.com/Theano/Theano/pull/4747
-* https://github.com/Theano/Theano/pull/4743
-* https://github.com/Theano/Theano/pull/4737
-* https://github.com/Theano/Theano/pull/4733
-* https://github.com/Theano/Theano/pull/4725
-* https://github.com/Theano/Theano/pull/4726
-* https://github.com/Theano/Theano/pull/4683
-* https://github.com/Theano/Theano/pull/4570
-* https://github.com/Theano/Theano/pull/4712
-* https://github.com/Theano/Theano/pull/4716
-* https://github.com/Theano/Theano/pull/4654
-* https://github.com/Theano/Theano/pull/4694
-* https://github.com/Theano/Theano/pull/4691
-* https://github.com/Theano/Theano/pull/4696
-* https://github.com/Theano/Theano/pull/4591
-* https://github.com/Theano/Theano/pull/4701
-* https://github.com/Theano/Theano/pull/4678
-* https://github.com/Theano/Theano/pull/4587
-* https://github.com/Theano/Theano/pull/4676
-* https://github.com/Theano/Theano/pull/4674
-* https://github.com/Theano/Theano/pull/4615
-* https://github.com/Theano/Theano/pull/4582
-* https://github.com/Theano/Theano/pull/4663
-* https://github.com/Theano/Theano/pull/4584
-* https://github.com/Theano/Theano/pull/4604
-* https://github.com/Theano/Theano/pull/4661
-* https://github.com/Theano/Theano/pull/4631
-* https://github.com/Theano/Theano/pull/4628
-* https://github.com/Theano/Theano/pull/4535
-* https://github.com/Theano/Theano/pull/4642
-* https://github.com/Theano/Theano/pull/4629
-* https://github.com/Theano/Theano/pull/4632
-* https://github.com/Theano/Theano/pull/4388
-* https://github.com/Theano/Theano/pull/4634
-* https://github.com/Theano/Theano/pull/4630
-* https://github.com/Theano/Theano/pull/4619
-* https://github.com/Theano/Theano/pull/4618
-* https://github.com/Theano/Theano/pull/4621
-* https://github.com/Theano/Theano/pull/4611
-* https://github.com/Theano/Theano/pull/4607
-* https://github.com/Theano/Theano/pull/4617
-* https://github.com/Theano/Theano/pull/4376
-* https://github.com/Theano/Theano/pull/4605
-* https://github.com/Theano/Theano/pull/4613
-* https://github.com/Theano/Theano/pull/4576
-* https://github.com/Theano/Theano/pull/4614
-* https://github.com/Theano/Theano/pull/4556
-* https://github.com/Theano/Theano/pull/4128
-* https://github.com/Theano/Theano/pull/4606
-* https://github.com/Theano/Theano/pull/4608
-* https://github.com/Theano/Theano/pull/4598
-* https://github.com/Theano/Theano/pull/4404
-* https://github.com/Theano/Theano/pull/4600
-* https://github.com/Theano/Theano/pull/4153
-* https://github.com/Theano/Theano/pull/4593
-* https://github.com/Theano/Theano/pull/4440
-* https://github.com/Theano/Theano/pull/4578
-* https://github.com/Theano/Theano/pull/4577
-* https://github.com/Theano/Theano/pull/4567
-* https://github.com/Theano/Theano/pull/4148
-* https://github.com/Theano/Theano/pull/4524
-* https://github.com/Theano/Theano/pull/4261
-* https://github.com/Theano/Theano/pull/4571
-* https://github.com/Theano/Theano/pull/4566
-* https://github.com/Theano/Theano/pull/4565
-* https://github.com/Theano/Theano/pull/4500
-* https://github.com/Theano/Theano/pull/4554
-* https://github.com/Theano/Theano/pull/4544
-* https://github.com/Theano/Theano/pull/4546
-* https://github.com/Theano/Theano/pull/4542
-* https://github.com/Theano/Theano/pull/4549
-* https://github.com/Theano/Theano/pull/4541
-* https://github.com/Theano/Theano/pull/4532
-* https://github.com/Theano/Theano/pull/4540
-* https://github.com/Theano/Theano/pull/4537
-* https://github.com/Theano/Theano/pull/4525
-* https://github.com/Theano/Theano/pull/4533
-* https://github.com/Theano/Theano/pull/4530
-* https://github.com/Theano/Theano/pull/4526
-* https://github.com/Theano/Theano/pull/4522
-* https://github.com/Theano/Theano/pull/4501
-* https://github.com/Theano/Theano/pull/4512
-* https://github.com/Theano/Theano/pull/4518
-* https://github.com/Theano/Theano/pull/4468
-* https://github.com/Theano/Theano/pull/4506
-* https://github.com/Theano/Theano/pull/4515
-* https://github.com/Theano/Theano/pull/4511
-* https://github.com/Theano/Theano/pull/4505
-* https://github.com/Theano/Theano/pull/4504
-* https://github.com/Theano/Theano/pull/4503
-* https://github.com/Theano/Theano/pull/4496
-* https://github.com/Theano/Theano/pull/4498
-* https://github.com/Theano/Theano/pull/4493
-* https://github.com/Theano/Theano/pull/4463
-* https://github.com/Theano/Theano/pull/4488
-* https://github.com/Theano/Theano/pull/4454
-* https://github.com/Theano/Theano/pull/4492
-* https://github.com/Theano/Theano/pull/4489
-* https://github.com/Theano/Theano/pull/4491
-* https://github.com/Theano/Theano/pull/4478
-* https://github.com/Theano/Theano/pull/4439
-* https://github.com/Theano/Theano/pull/4484
-* https://github.com/Theano/Theano/pull/4420
-* https://github.com/Theano/Theano/pull/4464
-* https://github.com/Theano/Theano/pull/4472
-* https://github.com/Theano/Theano/pull/4467
-* https://github.com/Theano/Theano/pull/4470
-* https://github.com/Theano/Theano/pull/4475
-* https://github.com/Theano/Theano/pull/4473
-* https://github.com/Theano/Theano/pull/4244
-* https://github.com/Theano/Theano/pull/4443
-* https://github.com/Theano/Theano/pull/4450
-* https://github.com/Theano/Theano/pull/4378
-* https://github.com/Theano/Theano/pull/4422
-* https://github.com/Theano/Theano/pull/4413
-* https://github.com/Theano/Theano/pull/4418
-* https://github.com/Theano/Theano/pull/4419
-* https://github.com/Theano/Theano/pull/4383
-* https://github.com/Theano/Theano/pull/4062
-* https://github.com/Theano/Theano/pull/4410
-* https://github.com/Theano/Theano/pull/4343
-* https://github.com/Theano/Theano/pull/4415
-* https://github.com/Theano/Theano/pull/4414
-* https://github.com/Theano/Theano/pull/4401
-* https://github.com/Theano/Theano/pull/4367
-* https://github.com/Theano/Theano/pull/4406
-* https://github.com/Theano/Theano/pull/4408
-* https://github.com/Theano/Theano/pull/4407
-* https://github.com/Theano/Theano/pull/4323
-* https://github.com/Theano/Theano/pull/4371
-* https://github.com/Theano/Theano/pull/4370
-* https://github.com/Theano/Theano/pull/4279
-* https://github.com/Theano/Theano/pull/4381
-* https://github.com/Theano/Theano/pull/4263
-* https://github.com/Theano/Theano/pull/4375
-* https://github.com/Theano/Theano/pull/4374
-* https://github.com/Theano/Theano/pull/4290
-* https://github.com/Theano/Theano/pull/4373
-* https://github.com/Theano/Theano/pull/4366
-* https://github.com/Theano/Theano/pull/4362
-* https://github.com/Theano/Theano/pull/4363
-* https://github.com/Theano/Theano/pull/4344
-* https://github.com/Theano/Theano/pull/4335
-* https://github.com/Theano/Theano/pull/4330
-* https://github.com/Theano/Theano/pull/4324
-* https://github.com/Theano/Theano/pull/4340
-* https://github.com/Theano/Theano/pull/4336
-* https://github.com/Theano/Theano/pull/4327
-* https://github.com/Theano/Theano/pull/4333
-* https://github.com/Theano/Theano/pull/4301
-* https://github.com/Theano/Theano/pull/4310
-* https://github.com/Theano/Theano/pull/4309
-* https://github.com/Theano/Theano/pull/4298
-* https://github.com/Theano/Theano/pull/4256
-* https://github.com/Theano/Theano/pull/4287
-* https://github.com/Theano/Theano/pull/4288
-* https://github.com/Theano/Theano/pull/4286
-* https://github.com/Theano/Theano/pull/3983
-* https://github.com/Theano/Theano/pull/4246
-* https://github.com/Theano/Theano/pull/3758
-* https://github.com/Theano/Theano/pull/4285
-* https://github.com/Theano/Theano/pull/4278
-* https://github.com/Theano/Theano/pull/4216
-* https://github.com/Theano/Theano/pull/4273
-* https://github.com/Theano/Theano/pull/4281
-* https://github.com/Theano/Theano/pull/4245
-* https://github.com/Theano/Theano/pull/4250
-* https://github.com/Theano/Theano/pull/4235
-* https://github.com/Theano/Theano/pull/4260
+* https://github.com/Theano/Theano/pull/5715
+* https://github.com/Theano/Theano/pull/5502
+* https://github.com/Theano/Theano/pull/5533
+* https://github.com/Theano/Theano/pull/5660
+* https://github.com/Theano/Theano/pull/5682
+* https://github.com/Theano/Theano/pull/5704
+* https://github.com/Theano/Theano/pull/5687
+* https://github.com/Theano/Theano/pull/5455
+* https://github.com/Theano/Theano/pull/5667
+* https://github.com/Theano/Theano/pull/5554
+* https://github.com/Theano/Theano/pull/5486
+* https://github.com/Theano/Theano/pull/5567
+* https://github.com/Theano/Theano/pull/5615
+* https://github.com/Theano/Theano/pull/5672
+* https://github.com/Theano/Theano/pull/5524

 Theano Development version
 ==========================

--- a/doc/conf.py
+++ b/doc/conf.py
@@ -74,7 +74,7 @@ copyright = '2008--2017, LISA lab'
 # The short X.Y version.
 version = '0.9'
 # The full version, including alpha/beta/rc tags.
-release = '0.9.0rc4'
+release = '0.9.0'

 # There are two options for replacing |today|: either, you set today to some
 # non-false value, then it is used:

--- a/doc/index.txt
+++ b/doc/index.txt
@@ -21,6 +21,8 @@ learning/machine learning <https://mila.umontreal.ca/en/cours/>`_ classes).
 News
 ====

+* 2017/03/20: Release of Theano 0.9.0. Everybody is encouraged to update.
+
 * 2017/03/13: Release of Theano 0.9.0rc4, with crash fixes and bug fixes.

 * 2017/03/06: Release of Theano 0.9.0rc3, with crash fixes, bug fixes and improvements.

--- a/doc/internal/how_to_release.txt
+++ b/doc/internal/how_to_release.txt
@@ -33,7 +33,7 @@ Edit ``setup.py`` to contain the newest version number ::
    cd Theano-0.X
    vi setup.py     # Edit the MAJOR, MINOR, MICRO and SUFFIX

-``conf.py`` in the ``doc/`` directory should be updated in the following ways:
+``Theano/doc/conf.py`` should be updated in the following ways:

 * Change the ``version`` and ``release`` variables to new version number.
 * Change the upper copyright year to the current year if necessary.

--- a/doc/introduction.txt
+++ b/doc/introduction.txt
@@ -165,7 +165,7 @@ Note: There is no short term plan to support multi-node computation.
 Theano Vision State
 ===================

-Here is the state of that vision as of March 13th, 2017 (after Theano 0.9.0rc4):
+Here is the state of that vision as of March 20th, 2017 (after Theano 0.9.0):

 * We support tensors using the `numpy.ndarray` object and we support many operations on them.
 * We support sparse types by using the `scipy.{csc,csr,bsr}_matrix` object and support some operations on them.

--- a/setup.py
+++ b/setup.py
@@ -53,7 +53,7 @@ PLATFORMS           = ["Windows", "Linux", "Solaris", "Mac OS-X", "Unix"]
 MAJOR               = 0
 MINOR               = 9
 MICRO               = 0
-SUFFIX              = "rc4"  # Should be blank except for rc's, betas, etc.
+SUFFIX              = ""  # Should be blank except for rc's, betas, etc.
 ISRELEASED          = False

 VERSION             = '%d.%d.%d%s' % (MAJOR, MINOR, MICRO, SUFFIX)