Skip to content

P
pytensor
提交 f46e5153 authored 作者: notoraptor's avatar notoraptor 提交者： GitHub
浏览文件
操作

Merge pull request #6279 from notoraptor/prepare-release-0.10-beta

Prepare beta release 0.10.0beta1.
上级 63dc46bc a3775a9f
隐藏空白字符变更
内嵌 并排
正在显示 包含 733 行增加 和 240 行删除
.mailmap
浏览文件 @ f46e5153
......@@ -15,6 +15,7 @@ abalkin <abalkin@enlnt.com> abalkin <abalkin>
abalkin <abalkin@enlnt.com> abalkin <serpent.speak@gmail.com>
abalkin <abalkin@enlnt.com> Alexander Belopolsky <abalkin@enlnt.com>
abalkin <abalkin@enlnt.com> Alexander Belopolsky <a@enlnt.com>
Aleksandar Botev <botevmg@gmail.com> botev <botevmg@gmail.com>
Alex Lamb <alex6200@gmail.com> AlexLamb <alex6200@gmail.com>
Alex Lamb <alex6200@gmail.com> DeathMonster666 <alex6200@gmail.com>
Alexandre de Brebisson <adbrebs@gmail.com> AdeB <adbrebs@gmail.com>
......@@ -25,6 +26,7 @@ Andre Holzner <Andre.Georg.Holzner@cern.ch> andreh <andreh@localhost>
Andre Holzner <Andre.Georg.Holzner@cern.ch> Andre Holzner <holzner@andres-mbp-2.fritz.box>
Arjun Jain <arjunjain@gmail.com> Arjun Jain <stencilman@users.noreply.github.com>
Arnaud Bergeron <abergeron@gmail.com> <abergeron@gmail.com>
Arnaud Bergeron <abergeron@gmail.com> <bergearn@iro.umontreal.ca>
<abergeron@gmail.com> <anakha@kami.(none)>
Balázs Hidasi <hidasi.balazs@gravityrd.com> Balázs <hidasib@gmail.com>
Bart van Merrienboer <bart.vanmerrienboer@gmail.com> Bart van Merriënboer <bart.vanmerrienboer@gmail.com>
......@@ -62,9 +64,14 @@ Dzmitry Bahdanau <dimabgv@gmail.com> rizar <dimabv@tut.by>
Eric Hunsberger <hunse@ctn> hunse <hunse@ctn>
Ethan Buchman <ebuchman@uoguelph.ca> ebuchman <ebuchman@uoguelph.ca>
Evelyn Mitchell <efm-github@linsomniac.com> evelynmitchell <efm-github@linsomniac.com>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@bart7.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@elisa1.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@eos3.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@kepler2.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@kepler3.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@elisa1.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@ceylon.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@leto21.iro.umontreal.ca>
Faruk Ahmed <faruk.ahmed.91@gmail.com> Faruk Ahmed <ahmedfar@leto23.iro.umontreal.ca>
Fei Wang <fay96816@gmail.com> fay <fay96816@gmail.com>
Francesco Visin <fvisin@gmail.com> Francesco <fvisin@users.noreply.github.com>
Francesco Visin <fvisin@gmail.com> fvisin <fvisin@gmail.com>
......@@ -81,6 +88,7 @@ Frederic Bastien <nouiz@nouiz.org> Frederic <nouiz@nouiz.org>
Frederic Bastien <nouiz@nouiz.org> Frédéric Bastien <frederic.bastien@gmail.com>
Frederic Bastien <nouiz@nouiz.org> theano-bot <frederic.bastien.1@umontreal.ca>
Gennadiy Tupitsin <genichyar@genichyar.com> genichyar <genichyar@genichyar.com>
Ghislain Antony Vaillant <ghisvail@gmail.com> Ghislain Antony Vaillant <ghisvail@users.noreply.github.com>
Gokula Krishnan <gokul.uf@gmail.com> Gokul <gokul.uf@gmail.com>
Grégoire Mesnil <gregoire.mesnil@gmail.com> Grégoire <gregoire.mesnil@laposte.net>
Grégoire Mesnil <gregoire.mesnil@gmail.com> Grégoire <gregoire.mesnil@gmail.com>
......@@ -90,6 +98,7 @@ Guillaume Desjardins <guillaume.desjardins@gmail.com> desjagui@atchoum.iro.umont
Guillaume Desjardins <guillaume.desjardins@gmail.com> desjagui@opale.iro.umontreal.ca <desjagui@opale.iro.umontreal.ca>
Guillaume Desjardins <guillaume.desjardins@gmail.com> gdesjardins <devnull@localhost>
Guillaume Desjardins <guillaume.desjardins@gmail.com> tutorial/debug_faq.txt <devnull@localhost>
gw0 [http://gw.tnode.com/] <gw.2015@tnode.com> gw0 [http://gw.tnode.com/] <gw.2016@tnode.com>
Hani Almousli <hani.mousli@gmail.com> Hani <hani.mousli@gmail.com>
Hani Almousli <hani.mousli@gmail.com> HaniAlmousli <hani.mousli@gmail.com>
Huy Nguyen <huy@huyng.com> huyng <huy@huyng.com>
......@@ -118,6 +127,8 @@ Jeremiah Lowin <jlowin@lowindata.com> Jeremiah Lowin <jlowin@iHal.local>
Jeremie Tanguay <tanguaj@iro.umontreal.ca> Tanjay94 <you@yourdomain.example.com>
Jeremie Tanguay <tanguaj@iro.umontreal.ca> Jeremie Tanguay <tanguaj@bart4.iro.umontreal.ca>
Jesse Livezey <jesse.livezey@berkeley.edu> JesseLivezey <jesse.livezey@gmail.com>
João Victor Tozatti Risso <joaovictortr@gmail.com> João Victor Risso <joaovictor.risso@gmail.com>
João Victor Tozatti Risso <joaovictortr@gmail.com> João Victor Tozatti Risso <joaovictor.risso@gmail.com>
John Salvatier <jsalvatier@gmail.com> jsalvatier <jsalvatier@gmail.com>
John Salvatier <jsalvatier@gmail.com> john salvatier <jsalvatier@gmail.com>
John Schulman <john.d.schulman@gmail.com> joschu <john.d.schulman@gmail.com>
......@@ -156,6 +167,7 @@ Markus Roth <markus.roth@herr-biber.de> Markus Roth <mail@rothmark.us>
Mathieu Germain <mathieu.germain@gmail.com> Mathieu Germain <mathieu.germain2@usherbrooke.ca>
Mehdi Mirza <memirzamo@gmail.com> Mehdi Mirza <memimo@users.noreply.github.com>
Mehdi Mirza <memirzamo@gmail.com> memimo <memirzamo@gmail.com>
Mohammed Affan <affanv14@gmail.com> affanv14 <affanv14@gmail.com>
Moslem Kazemi <moslemk@gmail.com> Moslem Kazemi <moslemk@users.noreply.github.com>
Moslem Kazemi <moslemk@gmail.com> Mo <moslemk@gmail.com>
Nan Rosemary Ke <rosemary.ke@west.cmu.edu> nke001 <rosemary.nan.ke@gmail.com>
......@@ -195,6 +207,7 @@ Razvan Pascanu <r.pascanu@gmail.com> Razvan Pascanu <rman@rman-Dell-System-XPS-L
Razvan Pascanu <r.pascanu@gmail.com> Razvan Pascanu <rman@rman-pad.(none)>
Razvan Pascanu <r.pascanu@gmail.com> pascanur@simplet.iro.umontreal.ca <pascanur@simplet.iro.umontreal.ca>
Razvan Pascanu <r.pascanu@gmail.com> rman@rpad <rman@rpad>
Reyhane Askari <r.askari.hemmat@gmail.com> Reyhane Askari <ReyhaneAskari@users.noreply.github.com>
Roy Xue <xljroy@gmail.com> Lijun Xue <xljroy@gmail.com>
Ruslana Makovetsky <ruslana@cim.mcgill.ca> ruslanagit <ruslana@cim.mcgill.ca>
Sander Dieleman <sanderdieleman@gmail.com> benanne <sanderdieleman@gmail.com>
......@@ -206,9 +219,14 @@ Simon Lefrancois <simon.lefrancois@umontreal.ca> slefrancois <simon.lefrancois@u
Simon Lefrancois <simon.lefrancois@umontreal.ca> Simon Lefrancois <lefransi@iro.umontreal.ca>
Sina Honari <honaris@iro.umontreal.ca> SinaHonari <sina2222@gmail.com>
Sina Honari <honaris@iro.umontreal.ca> Sina Honari <honaris@eos21.iro.umontreal.ca>
Sina Honari <honaris@iro.umontreal.ca> Sina Honari <sina.honari@gmail.com>
Søren Kaae Sønderby <skaaesonderby@gmail.com> skaae <skaaesonderby@gmail.com>
Steven Bocco <stevenbocco@gmail.com> notoraptor <stevenbocco@gmail.com>
Steven Bocco <stevenbocco@gmail.com> notoraptor <notoraptor@users.noreply.github.com>
Steven Bocco <stevenbocco@gmail.com> Seton Steven Bocco <boccoset@elisa1.iro.umontreal.ca>
Steven Bocco <stevenbocco@gmail.com> Seton Steven Bocco <boccoset@leto01.iro.umontreal.ca>
Steven Bocco <stevenbocco@gmail.com> Seton Steven Bocco <boccoset@leto15.iro.umontreal.ca>
Steven Bocco <stevenbocco@gmail.com> Seton Steven Bocco <boccoset@leto51.iro.umontreal.ca>
Steven Pigeon <pigeon@iro.umontreal.ca> steven-pigeon <pigeon@iro.umontreal.ca>
Thomas George <tfjgeorge@gmail.com> Thomas George <georgeth@helios1.helios>
Valentin Bisson <valentin.bisson@umontreal.ca> onze <onzeonline@gmail.com>
......
HISTORY.txt
浏览文件 @ f46e5153
......@@ -5,6 +5,254 @@
Old Release Notes
=================
Theano 0.9.0 (20th of March, 2017)
==================================
This is a final release of Theano, version ``0.9.0``, with a lot of
new features, interface changes, improvements and bug fixes.
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:
- 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:
- 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
- Graphs improvements:
- 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
- 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``
- Less C code compilation
- Added support for bool dtype
- Updated and more complete documentation
- Bug fixes related to merge optimizer and shape inference
- Lot of other bug fixes, crashes fixes and warning improvements
A total of 123 people contributed to this release since 0.8.0, see list below.
Interface changes:
- 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
- New pooling interface
- Pooling parameters can change at run time
- 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)
- Gemv(matrix-vector product) speed up for special shape
- 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:
- ``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, ...}``
- compiledir_format support device
- 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:
- 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
Other more detailed changes:
- 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
- Make scan gradient more deterministic
Commiters since 0.8.0:
- Frederic Bastien
- Arnaud Bergeron
- Pascal Lamblin
- Steven Bocco
- Ramana Subramanyam
- Simon Lefrancois
- Gijs van Tulder
- Benjamin Scellier
- khaotik
- Chiheb Trabelsi
- Chinnadhurai Sankar
- Cesar Laurent
- Reyhane Askari
- Mohammad Pezeshki
- Alexander Matyasko
- Alexandre de Brebisson
- Mathieu Germain
- Nan Rosemary Ke
- Pierre Luc Carrier
- Olivier Mastropietro
- Thomas George
- Saizheng Zhang
- Iulian Vlad Serban
- Francesco Visin
- Caglar
- Faruk Ahmed
- Harm de Vries
- Samira Shabanian
- Vincent Dumoulin
- Nicolas Ballas
- Jakub Sygnowski
- Jan Schlüter
- Samira Ebrahimi Kahou
- Mikhail Korobov
- Fei Wang
- Kv Manohar
- Jesse Livezey
- Kelvin Xu
- Matt Graham
- Ruslana Makovetsky
- Sina Honari
- Bryn Keller
- Ciyong Chen
- Vitaliy Kurlin
- Zhouhan LIN
- Gokula Krishnan
- Kumar Krishna Agrawal
- Ozan Çağlayan
- Vincent Michalski
- affanv14
- Amjad Almahairi
- Ray Donnelly
- Tim Cooijmans
- happygds
- mockingjamie
- 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
- p
- texot
- Andrés Gottlieb
- 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
- Kaixhin
- Kevin Keraudren
- Maltimore
- Marc-Alexandre Cote
- Marco
- Marius F. Killinger
- Martin Drawitsch
- Maxim Kochurov
- Micah Bojrab
- 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
- valtron
- wazeerzulfikar
- you-n-g
Theano 0.9.0rc4 (13th of March, 2017)
=====================================
......
NEWS.txt
浏览文件 @ f46e5153
......@@ -3,249 +3,182 @@ Release Notes
=============
Theano 0.9.0 (20th of March, 2017)
==================================
Theano 0.10.0beta1 (9th of August, 2017)
========================================
This is a final release of Theano, version ``0.9.0``, with a lot of
new features, interface changes, improvements and bug fixes.
This release contains a lot of bug fixes, improvements and new features to prepare the upcoming release candidate.
We recommend that everybody update to this version.
We recommend that every developer updates 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:
Highlights:
- Moved Python 3.* minimum supported version from 3.3 to 3.4
- Replaced deprecated package ``nose-parameterized`` with up-to-date package ``parameterized`` for Theano requirements
- Theano now internally uses ``sha256`` instead of ``md5`` to work on systems that forbide ``md5`` for security reason
- Removed old GPU backend ``theano.sandbox.cuda``. New backend ``theano.gpuarray`` is now the official GPU backend
- Support more debuggers for ``PdbBreakpoint``
- 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
- Scan improvements:
- Speed up Theano scan compilation and gradient computation
- Added meaningful message when missing inputs to scan
- 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
- Speed up graph toposort algorithm
- Faster C compilation by massively using a new interface for op params
- Faster optimization step
- Documentation updated and more complete
- Many bug fixes, crash fixes and warning improvements
- Graphs improvements:
- 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
- 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``
- Less C code compilation
- Added support for bool dtype
- Updated and more complete documentation
- Bug fixes related to merge optimizer and shape inference
- Lot of other bug fixes, crashes fixes and warning improvements
A total of 123 people contributed to this release since 0.8.0, see list below.
A total of 65 people contributed to this release since 0.9.0, see list below.
Interface changes:
- Merged ``CumsumOp/CumprodOp`` into ``CumOp``
- In MRG module:
- 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)
- Renamed ``MultinomialWOReplacementFromUniform`` to ``ChoiceFromUniform``
- Replaced method ``multinomial_wo_replacement()`` with new method ``choice()``
- Random generator now tries to infer the broadcast pattern of its output
- Removed or deprecated Theano flags:
- New pooling interface
- Pooling parameters can change at run time
- 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
- ``cublas.lib``
- ``cuda.enabled``
- ``enable_initial_driver_test``
- ``gpuarray.sync``
- ``home``
- ``lib.cnmem``
- ``nvcc.*`` flags
- ``pycuda.init``
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
- Changed ``grad()`` method to ``L_op()`` in ops that need the outputs to compute gradient
Convolution updates:
- Extended Theano flag ``dnn.enabled`` with new option ``no_check`` to help speed up cuDNN importation
- Implemented separable convolutions
- Implemented grouped convolutions
GPU:
- cuDNN: support versoin 5.1 and wrap batch normalization (2d and 3d) and RNN functions
- Multiple-GPU, synchrone update (via platoon, use NCCL)
- Gemv(matrix-vector product) speed up for special shape
- 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``
- Prevent GPU initialization when not required
- Added disk caching option for kernels
- Added method ``my_theano_function.sync_shared()`` to help synchronize GPU Theano functions
- Added useful stats for GPU in profile mode
- Added Cholesky op based on ``cusolver`` backend
- Added GPU ops based on `magma library <http://icl.cs.utk.edu/magma/software/>`_:
SVD, matrix inverse, QR, cholesky and eigh
- Added ``GpuCublasTriangularSolve``
- Added atomic addition and exchange for ``long long`` values in ``GpuAdvancedIncSubtensor1_dev20``
- Support log gamma function for all non-complex types
- Support GPU SoftMax in both OpenCL and CUDA
- Support offset parameter ``k`` for ``GpuEye``
- ``CrossentropyCategorical1Hot`` and its gradient are now lifted to GPU
- Better cuDNN support
- Official support for ``v5.*`` and ``v6.*``
- Better support and loading on Windows and Mac
- Support cuDNN v6 dilated convolutions
- Support cuDNN v6 reductions
- Added new Theano flags ``cuda.include_path``, ``dnn.base_path`` and ``dnn.bin_path``
to help configure Theano when CUDA and cuDNN can not be found automatically.
- Updated ``float16`` support
- Added documentation for GPU float16 ops
- Support ``float16`` for ``GpuGemmBatch``
- Started to use ``float32`` precision for computations that don't support ``float16`` on GPU
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 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, ...}``
- compiledir_format support device
- 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)
- Added a wrapper for `Baidu's CTC <https://github.com/baidu-research/warp-ctc>`_ cost and gradient functions
- Added scalar and elemwise CPU ops for modified Bessel function of order 0 and 1 from ``scipy.special``.
- Added Scaled Exponential Linear Unit (SELU) activation
- Added sigmoid_binary_crossentropy function
- Added tri-gamma function
- Added modes ``half`` and ``full`` for ``Images2Neibs`` ops
- Implemented gradient for ``AbstractBatchNormTrainGrad``
- Implemented gradient for matrix pseudoinverse op
- Added new prop `replace` for ``ChoiceFromUniform`` op
- Added new prop ``on_error`` for CPU ``Cholesky`` op
- Added new Theano flag ``deterministic`` to help control how Theano optimize certain ops that have deterministic versions.
Currently used for subtensor Ops only.
- Added new Theano flag ``cycle_detection`` to speed-up optimization step by reducing time spending in inplace optimizations
- Added new Theano flag ``check_stack_trace`` to help check the stack trace during optimization process
- Added new Theano flag ``cmodule.debug`` to allow a debug mode for Theano C code. Currently used for cuDNN convolutions only.
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)
- More stack trace in error messages
- Speed up cholesky grad
- ``log(sum(exp(...)))`` now get stability optimized
- Added deprecation warning for the softmax and logsoftmax vector case
- Added a warning to announce that C++ compiler will become mandatory in next Theano release ``0.11``
Other more detailed changes:
- 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
- Make scan gradient more deterministic
Commiters since 0.8.0:
- Removed useless warning when profile is manually disabled
- Added tests for abstract conv
- Added options for `disconnected_outputs` to Rop
- Removed ``theano/compat/six.py``
- Removed ``COp.get_op_params()``
- Support of list of strings for ``Op.c_support_code()``, to help not duplicate support codes
- Macro names provided for array properties are now standardized in both CPU and GPU C codes
- Started to move C code files into separate folder ``c_code`` in every Theano module
- Many improvements for Travis CI tests (with better splitting for faster testing)
- Many improvements for Jenkins CI tests: daily testings on Mac and Windows in addition to Linux
Commiters since 0.9.0:
- Frederic Bastien
- Arnaud Bergeron
- Pascal Lamblin
- amrithasuresh
- João Victor Tozatti Risso
- Steven Bocco
- Ramana Subramanyam
- Simon Lefrancois
- Gijs van Tulder
- Benjamin Scellier
- khaotik
- Chiheb Trabelsi
- Chinnadhurai Sankar
- Cesar Laurent
- Pascal Lamblin
- Mohammed Affan
- Reyhane Askari
- Mohammad Pezeshki
- Alexander Matyasko
- Alexandre de Brebisson
- Mathieu Germain
- Nan Rosemary Ke
- Pierre Luc Carrier
- Olivier Mastropietro
- Simon Lefrancois
- Shawn Tan
- Thomas George
- Saizheng Zhang
- Iulian Vlad Serban
- Francesco Visin
- Caglar
- Faruk Ahmed
- Harm de Vries
- Samira Shabanian
- Vincent Dumoulin
- Nicolas Ballas
- Jakub Sygnowski
- Jan Schlüter
- Samira Ebrahimi Kahou
- Mikhail Korobov
- Fei Wang
- Kv Manohar
- Jesse Livezey
- Kelvin Xu
- Matt Graham
- Ruslana Makovetsky
- Sina Honari
- Bryn Keller
- Ciyong Chen
- Vitaliy Kurlin
- Zhouhan LIN
- Gokula Krishnan
- Kumar Krishna Agrawal
- Ozan Çağlayan
- Vincent Michalski
- affanv14
- Amjad Almahairi
- Ray Donnelly
- Tim Cooijmans
- happygds
- mockingjamie
- Christos Tsirigotis
- Aleksandar Botev
- jhelie
- xiaoqie
- Tegan Maharaj
- Matt Graham
- Cesar Laurent
- Gabe Schwartz
- Juan Camilo Gamboa Higuera
- AndroidCloud
- Saizheng Zhang
- vipulraheja
- 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
- Sina Honari
- Vikram
- erakra
- Chiheb Trabelsi
- Shubh Vachher
- Daren Eiri
- Gijs van Tulder
- Laurent Dinh
- Mohamed Ishmael Diwan Belghazi
- mila
- Jeff Donahue
- Ramana Subramanyam
- Bogdan Budescu
- Ghislain Antony Vaillant
- Jan Schlüter
- Xavier Bouthillier
- p
- texot
- Andrés Gottlieb
- 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
- Kaixhin
- Kevin Keraudren
- Maltimore
- Marc-Alexandre Cote
- Marco
- Marius F. Killinger
- Martin Drawitsch
- Maxim Kochurov
- Micah Bojrab
- 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
- valtron
- wazeerzulfikar
- you-n-g
- fo40225
- Aarni Koskela
- Adam Becker
- Adam Geitgey
- Adrian Keet
- Adrian Seyboldt
- Andrei Costinescu
- Anmol Sahoo
- Chong Wu
- Holger Kohr
- Jayanth Koushik
- Jenkins
- Lilian Besson
- Lv Tao
- Michael Manukyan
- Murugesh Marvel
- NALEPA
- Ubuntu
- Zotov Yuriy
- dareneiri
- lrast
- morrme
- yikang
NEWS_DEV.txt
浏览文件 @ f46e5153
......@@ -11,51 +11,343 @@ git shortlog -sn rel-0.9.0..
# docker?
# docker added to jenkins buildbot
TODO: better Theano conv doc
# NB: Following notes contains infos since 0.9.0.
Highlights:
- 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
- Moved Python 3.* minimum supported version from 3.3 to 3.4
- Replaced deprecated package ``nose-parameterized`` with up-to-date package ``parameterized`` for Theano requirements
- Theano now internally uses ``sha256`` instead of ``md5`` to work on systems that forbide ``md5`` for security reason
- Removed old GPU backend ``theano.sandbox.cuda``. New backend ``theano.gpuarray`` is now the official GPU backend
- Support more debuggers for ``PdbBreakpoint``
- New GPU back-end:
- Scan improvements
- Added useful stats for GPU in profile mode
- Added documentation for GPU float16 ops
- Speed up Theano scan compilation and gradient computation
- Added meaningful message when missing inputs to scan
- Speed up graph toposort algorithm
- Faster C compilation by massively using a new interface for op params
- Faster optimization step
- Documentation updated and more complete
- Many bug fixes, crash fixes and warning improvements
Interface changes:
- 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)
- Renamed ``MultinomialWOReplacementFromUniform`` to ``ChoiceFromUniform``
- Removed or deprecated Theano flags:
- ``cublas.lib``
- ``cuda.enabled``
- ``enable_initial_driver_test``
- ``gpuarray.sync``
- ``home``
- ``lib.cnmem``
- ``nvcc.*`` flags
- ``pycuda.init``
- Changed ``grad()`` method to ``L_op()`` in ops that need the outputs to compute gradient
Convolution updates:
- Extended Theano flag ``dnn.enabled`` with new option ``no_check`` to help speed up cuDNN importation
- Implemented separable convolutions
- Implemented grouped convolutions
GPU:
- ...
- Prevent GPU initialization when not required
- Added disk caching option for kernels
- Added method ``my_theano_function.sync_shared()`` to help synchronize GPU Theano functions
- Added useful stats for GPU in profile mode
- Added Cholesky op based on ``cusolver`` backend
- Added GPU ops based on `magma library <http://icl.cs.utk.edu/magma/software/>`_:
SVD, matrix inverse, QR, cholesky and eigh
- Added ``GpuCublasTriangularSolve``
- Added atomic addition and exchange for ``long long`` values in ``GpuAdvancedIncSubtensor1_dev20``
- Support log gamma function for all non-complex types
- Support GPU SoftMax in both OpenCL and CUDA
- Support offset parameter ``k`` for ``GpuEye``
- ``CrossentropyCategorical1Hot`` and its gradient are now lifted to GPU
New features:
- Added scalar and elemwise ops for modified Bessel function of order 0 and 1 from ``scipy.special``
- Better cuDNN support
- Official support for ``v5.*`` and ``v6.*``
- Better support and loading on Windows and Mac
- Support cuDNN v6 dilated convolutions
- Support cuDNN v6 reductions
- Added new Theano flags ``cuda.include_path``, ``dnn.base_path`` and ``dnn.bin_path``
to help configure Theano when CUDA and cuDNN can not be found automatically.
Others:
- ...
- Updated ``float16`` support
- Added documentation for GPU float16 ops
- Support ``float16`` for ``GpuGemmBatch``
- Started to use ``float32`` precision for computations that don't support ``float16`` on GPU
New features:
- Added a wrapper for `Baidu's CTC <https://github.com/baidu-research/warp-ctc>`_ cost and gradient functions
- Added scalar and elemwise CPU ops for modified Bessel function of order 0 and 1 from ``scipy.special``.
- Added Scaled Exponential Linear Unit (SELU) activation
- Added sigmoid_binary_crossentropy function
- Added tri-gamma function
- Added modes ``half`` and ``full`` for ``Images2Neibs`` ops
- Implemented gradient for ``AbstractBatchNormTrainGrad``
- Implemented gradient for matrix pseudoinverse op
- Added new prop `replace` for ``ChoiceFromUniform`` op
- Added new prop ``on_error`` for CPU ``Cholesky`` op
- Added new Theano flag ``deterministic`` to help control how Theano optimize certain ops that have deterministic versions.
Currently used for subtensor Ops only.
- Added new Theano flag ``cycle_detection`` to speed-up optimization step by reducing time spending in inplace optimizations
- Added new Theano flag ``check_stack_trace`` to help check the stack trace during optimization process
- Added new Theano flag ``cmodule.debug`` to allow a debug mode for Theano C code. Currently used for cuDNN convolutions only.
Others:
- Added deprecation warning for the softmax and logsoftmax vector case
- Added a warning to announce that C++ compiler will become mandatory in next Theano release ``0.11``
Other more detailed changes:
- Removed useless warning when profile is manually disabled
- Added tests for abstract conv
- Added options for `disconnected_outputs` to Rop
- Removed ``theano/compat/six.py``
- Removed ``COp.get_op_params()``
- Support of list of strings for ``Op.c_support_code()``, to help not duplicate support codes
- Macro names provided for array properties are now standardized in both CPU and GPU C codes
- Started to move C code files into separate folder ``c_code`` in every Theano module
- Many improvements for Travis CI tests (with better splitting for faster testing)
- Many improvements for Jenkins CI tests: daily testings on Mac and Windows in addition to Linux
ALL THE PR BELLOW HAVE BEEN CHECKED
* https://github.com/Theano/Theano/pull/6275
* https://github.com/Theano/Theano/pull/6218
* https://github.com/Theano/Theano/pull/6271
* https://github.com/Theano/Theano/pull/6253
* https://github.com/Theano/Theano/pull/6273
* https://github.com/Theano/Theano/pull/6262
* https://github.com/Theano/Theano/pull/6214
* https://github.com/Theano/Theano/pull/6264
* https://github.com/Theano/Theano/pull/6256
* https://github.com/Theano/Theano/pull/6254
* https://github.com/Theano/Theano/pull/6220
* https://github.com/Theano/Theano/pull/5949
* https://github.com/Theano/Theano/pull/6243
* https://github.com/Theano/Theano/pull/6250
* https://github.com/Theano/Theano/pull/6225
* https://github.com/Theano/Theano/pull/6242
* https://github.com/Theano/Theano/pull/6213
* https://github.com/Theano/Theano/pull/6199
* https://github.com/Theano/Theano/pull/6209
* https://github.com/Theano/Theano/pull/6216
* https://github.com/Theano/Theano/pull/6215
* https://github.com/Theano/Theano/pull/6182
* https://github.com/Theano/Theano/pull/6194
* https://github.com/Theano/Theano/pull/6190
* https://github.com/Theano/Theano/pull/6146
* https://github.com/Theano/Theano/pull/6201
* https://github.com/Theano/Theano/pull/6150
* https://github.com/Theano/Theano/pull/6204
* https://github.com/Theano/Theano/pull/6166
* https://github.com/Theano/Theano/pull/6174
* https://github.com/Theano/Theano/pull/6205
* https://github.com/Theano/Theano/pull/6183
* https://github.com/Theano/Theano/pull/6186
* https://github.com/Theano/Theano/pull/6203
* https://github.com/Theano/Theano/pull/6161
* https://github.com/Theano/Theano/pull/6164
* https://github.com/Theano/Theano/pull/6050
* https://github.com/Theano/Theano/pull/6178
* https://github.com/Theano/Theano/pull/6180
* https://github.com/Theano/Theano/pull/6173
* https://github.com/Theano/Theano/pull/6170
* https://github.com/Theano/Theano/pull/6092
* https://github.com/Theano/Theano/pull/6163
* https://github.com/Theano/Theano/pull/6171
* https://github.com/Theano/Theano/pull/6169
* https://github.com/Theano/Theano/pull/6165
* https://github.com/Theano/Theano/pull/5914
* https://github.com/Theano/Theano/pull/5775
* https://github.com/Theano/Theano/pull/6147
* https://github.com/Theano/Theano/pull/6159
* https://github.com/Theano/Theano/pull/6156
* https://github.com/Theano/Theano/pull/6154
* https://github.com/Theano/Theano/pull/5991
* https://github.com/Theano/Theano/pull/6149
* https://github.com/Theano/Theano/pull/6151
* https://github.com/Theano/Theano/pull/6116
* https://github.com/Theano/Theano/pull/6111
* https://github.com/Theano/Theano/pull/6139
* https://github.com/Theano/Theano/pull/6097
* https://github.com/Theano/Theano/pull/6070
* https://github.com/Theano/Theano/pull/6148
* https://github.com/Theano/Theano/pull/6140
* https://github.com/Theano/Theano/pull/6138
* https://github.com/Theano/Theano/pull/5881
* https://github.com/Theano/Theano/pull/6130
* https://github.com/Theano/Theano/pull/6044
* https://github.com/Theano/Theano/pull/6060
* https://github.com/Theano/Theano/pull/6109
* https://github.com/Theano/Theano/pull/6119
* https://github.com/Theano/Theano/pull/6123
* https://github.com/Theano/Theano/pull/6117
* https://github.com/Theano/Theano/pull/6120
* https://github.com/Theano/Theano/pull/5747
* https://github.com/Theano/Theano/pull/6087
* https://github.com/Theano/Theano/pull/6108
* https://github.com/Theano/Theano/pull/6112
* https://github.com/Theano/Theano/pull/6106
* https://github.com/Theano/Theano/pull/6107
* https://github.com/Theano/Theano/pull/6105
* https://github.com/Theano/Theano/pull/6102
* https://github.com/Theano/Theano/pull/6101
* https://github.com/Theano/Theano/pull/6077
* https://github.com/Theano/Theano/pull/6085
* https://github.com/Theano/Theano/pull/6091
* https://github.com/Theano/Theano/pull/6013
* https://github.com/Theano/Theano/pull/6088
* https://github.com/Theano/Theano/pull/6069
* https://github.com/Theano/Theano/pull/6084
* https://github.com/Theano/Theano/pull/6083
* https://github.com/Theano/Theano/pull/6081
* https://github.com/Theano/Theano/pull/6072
* https://github.com/Theano/Theano/pull/6045
* https://github.com/Theano/Theano/pull/6082
* https://github.com/Theano/Theano/pull/6049
* https://github.com/Theano/Theano/pull/6076
* https://github.com/Theano/Theano/pull/6062
* https://github.com/Theano/Theano/pull/6041
* https://github.com/Theano/Theano/pull/6057
* https://github.com/Theano/Theano/pull/6055
* https://github.com/Theano/Theano/pull/6056
* https://github.com/Theano/Theano/pull/6043
* https://github.com/Theano/Theano/pull/6032
* https://github.com/Theano/Theano/pull/6030
* https://github.com/Theano/Theano/pull/5942
* https://github.com/Theano/Theano/pull/6025
* https://github.com/Theano/Theano/pull/6038
* https://github.com/Theano/Theano/pull/6034
* https://github.com/Theano/Theano/pull/6012
* https://github.com/Theano/Theano/pull/6029
* https://github.com/Theano/Theano/pull/6015
* https://github.com/Theano/Theano/pull/6027
* https://github.com/Theano/Theano/pull/6026
* https://github.com/Theano/Theano/pull/5980
* https://github.com/Theano/Theano/pull/6021
* https://github.com/Theano/Theano/pull/6022
* https://github.com/Theano/Theano/pull/6011
* https://github.com/Theano/Theano/pull/5935
* https://github.com/Theano/Theano/pull/5955
* https://github.com/Theano/Theano/pull/6009
* https://github.com/Theano/Theano/pull/5016
* https://github.com/Theano/Theano/pull/5794
* https://github.com/Theano/Theano/pull/5996
* https://github.com/Theano/Theano/pull/5923
* https://github.com/Theano/Theano/pull/5993
* https://github.com/Theano/Theano/pull/5983
* https://github.com/Theano/Theano/pull/5964
* https://github.com/Theano/Theano/pull/5940
* https://github.com/Theano/Theano/pull/5915
* https://github.com/Theano/Theano/pull/5989
* https://github.com/Theano/Theano/pull/5988
* https://github.com/Theano/Theano/pull/5987
* https://github.com/Theano/Theano/pull/5908
* https://github.com/Theano/Theano/pull/5974
* https://github.com/Theano/Theano/pull/5965
* https://github.com/Theano/Theano/pull/5960
* https://github.com/Theano/Theano/pull/5957
* https://github.com/Theano/Theano/pull/5936
* https://github.com/Theano/Theano/pull/5950
* https://github.com/Theano/Theano/pull/5948
* https://github.com/Theano/Theano/pull/5946
* https://github.com/Theano/Theano/pull/5947
* https://github.com/Theano/Theano/pull/5927
* https://github.com/Theano/Theano/pull/5944
* https://github.com/Theano/Theano/pull/5918
* https://github.com/Theano/Theano/pull/5941
* https://github.com/Theano/Theano/pull/5931
* https://github.com/Theano/Theano/pull/5937
* https://github.com/Theano/Theano/pull/5852
* https://github.com/Theano/Theano/pull/5922
* https://github.com/Theano/Theano/pull/5921
* https://github.com/Theano/Theano/pull/5902
* https://github.com/Theano/Theano/pull/5903
* https://github.com/Theano/Theano/pull/5909
* https://github.com/Theano/Theano/pull/5758
* https://github.com/Theano/Theano/pull/5778
* https://github.com/Theano/Theano/pull/5900
* https://github.com/Theano/Theano/pull/5895
* https://github.com/Theano/Theano/pull/5883
* https://github.com/Theano/Theano/pull/5896
* https://github.com/Theano/Theano/pull/5888
* https://github.com/Theano/Theano/pull/5886
* https://github.com/Theano/Theano/pull/5885
* https://github.com/Theano/Theano/pull/5873
* https://github.com/Theano/Theano/pull/5877
* https://github.com/Theano/Theano/pull/5878
* https://github.com/Theano/Theano/pull/5872
* https://github.com/Theano/Theano/pull/5870
* https://github.com/Theano/Theano/pull/5854
* https://github.com/Theano/Theano/pull/5865
* https://github.com/Theano/Theano/pull/5853
* https://github.com/Theano/Theano/pull/5850
* https://github.com/Theano/Theano/pull/5538
* https://github.com/Theano/Theano/pull/5863
* https://github.com/Theano/Theano/pull/5799
* https://github.com/Theano/Theano/pull/5859
* https://github.com/Theano/Theano/pull/5755
* https://github.com/Theano/Theano/pull/5860
* https://github.com/Theano/Theano/pull/5716
* https://github.com/Theano/Theano/pull/5842
* https://github.com/Theano/Theano/pull/5821
* https://github.com/Theano/Theano/pull/5789
* https://github.com/Theano/Theano/pull/5847
* https://github.com/Theano/Theano/pull/5735
* https://github.com/Theano/Theano/pull/5710
* https://github.com/Theano/Theano/pull/5843
* https://github.com/Theano/Theano/pull/5832
* https://github.com/Theano/Theano/pull/5814
* https://github.com/Theano/Theano/pull/5835
* https://github.com/Theano/Theano/pull/5834
* https://github.com/Theano/Theano/pull/5829
* https://github.com/Theano/Theano/pull/5785
* https://github.com/Theano/Theano/pull/5824
* https://github.com/Theano/Theano/pull/5820
* https://github.com/Theano/Theano/pull/5808
* https://github.com/Theano/Theano/pull/5815
* https://github.com/Theano/Theano/pull/5819
* https://github.com/Theano/Theano/pull/5612
* https://github.com/Theano/Theano/pull/5802
* https://github.com/Theano/Theano/pull/5796
* https://github.com/Theano/Theano/pull/5806
* https://github.com/Theano/Theano/pull/5782
* https://github.com/Theano/Theano/pull/5787
* https://github.com/Theano/Theano/pull/5774
* https://github.com/Theano/Theano/pull/5751
* https://github.com/Theano/Theano/pull/5779
* https://github.com/Theano/Theano/pull/5763
* https://github.com/Theano/Theano/pull/5746
* https://github.com/Theano/Theano/pull/5579
* https://github.com/Theano/Theano/pull/5772
* https://github.com/Theano/Theano/pull/5756
* https://github.com/Theano/Theano/pull/5769
* https://github.com/Theano/Theano/pull/5433
* https://github.com/Theano/Theano/pull/5760
* https://github.com/Theano/Theano/pull/5470
* https://github.com/Theano/Theano/pull/5759
* https://github.com/Theano/Theano/pull/5739
* https://github.com/Theano/Theano/pull/5752
* https://github.com/Theano/Theano/pull/5548
* https://github.com/Theano/Theano/pull/5749
* https://github.com/Theano/Theano/pull/5665
* https://github.com/Theano/Theano/pull/5562
* https://github.com/Theano/Theano/pull/5686
* https://github.com/Theano/Theano/pull/5718
* https://github.com/Theano/Theano/pull/5698
* https://github.com/Theano/Theano/pull/5720
* https://github.com/Theano/Theano/pull/5717
* https://github.com/Theano/Theano/pull/5715
* https://github.com/Theano/Theano/pull/5502
* https://github.com/Theano/Theano/pull/5533
......
doc/conf.py
浏览文件 @ f46e5153
......@@ -72,9 +72,9 @@ copyright = '2008--2017, LISA lab'
# other places throughout the built documents.
#
# The short X.Y version.
version = '0.9'
version = '0.10'
# The full version, including alpha/beta/rc tags.
release = '0.9.0'
release = '0.10.0beta1'
# There are two options for replacing |today|: either, you set today to some
# non-false value, then it is used:
......
doc/index.txt
浏览文件 @ f46e5153
......@@ -21,6 +21,8 @@ learning/machine learning <https://mila.umontreal.ca/en/cours/>`_ classes).
News
====
* 2017/08/09: Release of Theano 0.10.0beta1, many improvements and bugfixes, release candidate to coming.
* Removed support for the old (device=gpu) backend. Use the new
backend (device=cuda) for gpu computing. See `Converting to the new
gpu back end(gpuarray)
......
doc/install_generic.inc
浏览文件 @ f46e5153
......@@ -56,12 +56,12 @@ libgpuarray
:::::::::::
For the stable version of Theano you need a specific version of libgpuarray,
that has been tagged ``v0.6.5``.
that has been tagged ``v0.6.9``.
Download it with::
git clone https://github.com/Theano/libgpuarray.git
cd libgpuarray
git checkout tags/v0.6.5 -b v0.6.5
git checkout tags/v0.6.5 -b v0.6.9
and then follow the `Step-by-step instructions <http://deeplearning.net/software/libgpuarray/installation.html#step-by-step-install>`__.
......
doc/introduction.txt
浏览文件 @ f46e5153
......@@ -165,11 +165,11 @@ 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 20th, 2017 (after Theano 0.9.0):
Here is the state of that vision as of August 9th, 2017 (after Theano 0.10.0beta1):
* 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.
* We have implementing/wrapping more advanced linear algebra operations. Still many more possible.
* We have implementing/wrapping more advanced linear algebra operations. Still more possible.
* We have basic support for the creation of new operations from graphs at runtime. It supports well gradient overload
for every input and inlining at the start of compilation. We don't cover well the case when it is not inlined.
* We have many graph transformations that cover the 4 categories listed above.
......
setup.py
浏览文件 @ f46e5153
......@@ -53,7 +53,7 @@ PLATFORMS = ["Windows", "Linux", "Solaris", "Mac OS-X", "Unix"]
MAJOR = 0
MINOR = 10
MICRO = 0
SUFFIX = "dev1" # Should be blank except for rc's, betas, etc.
SUFFIX = "beta1" # Should be blank except for rc's, betas, etc.
ISRELEASED = False
VERSION = '%d.%d.%d%s' % (MAJOR, MINOR, MICRO, SUFFIX)
......
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论