hash
stringlengths
40
40
msg
stringlengths
1
131k
author
stringlengths
1
33
email
stringlengths
0
57
date
int64
1,447B
1,698B
eff1493fb96e6e565b47c0bb83c18098779dd8bb
Force copy of indices to local storage always. Change: 116898271
A. Unique TensorFlower
nobody@tensorflow.org
1,457,642,944,000
fd839dc6e22fd5e4766db237808828a9ff01a19c
Make sparse_softmax_cross_entropy_with_logits handle int32 labels Fixes #1250. Change: 116899325
Geoffrey Irving
geoffreyi@google.com
1,457,643,529,000
7cc6ba76176b84c828dd6b40ec5d2bc0d481f46a
Improve the BiasGrad for NCHW using less shared memory and better memory efficiency. With GoogleNet V1, time spent in BiasGrad in ms: Before After Improvement GoogleNet V1 19.70 13.14 49.93% Change: 116901889
Xiaoqiang Zheng
zhengxq@google.com
1,457,645,013,000
6d05cdb80b089ccbf05151101eb4943ffbbb529d
Ensure that tf-image-dashboard always sorts both runs and tags. Change: 116904157
Dan Mané
danmane@gmail.com
1,457,646,348,000
5b5976a3854776eb41494bd4b194a096ef3b28d0
Move all gRPC deps from WORKSPACE to workspace.bzl. Change: 116906391
Kiril Gorovoy
kgorovoy@google.com
1,457,647,568,000
e71090a1190d039b124ecfddd90f901e328647ea
Update Bower dependencies to latest G3 versions. Change: 116908297
Dan Mané
danmane@gmail.com
1,457,648,547,000
025c0d21a6689f081082b1a51f8812e56b07af77
Improve the loading of the variables for freeze_graph Change: 116910206
Jianmin Chen
goog.jmchen@gmail.com
1,457,649,588,000
8746ebaa91740ffbf96bd82279f831c6b2ca4307
Updated imports in skflow and examples, added BUILD
Illia Polosukhin
ilblackdragon@gmail.com
1,457,657,243,000
23bfde3a01ec051c59e2029dd66c9cc71dc17b1b
Updated readme for skflow examples to explain more about it and installation requirements
Illia Polosukhin
ilblackdragon@gmail.com
1,457,658,173,000
88682ce7e71d7ba5cf0764e3266fcac5dfa01f94
skflow: Added tests to BUILD and updated tests to import skflow via tensorflow
Illia Polosukhin
ilblackdragon@gmail.com
1,457,660,326,000
99c01169df045b50eeaddaa81b332b30d965942c
skflow: Added scipy and sklearn package installation to CI
Illia Polosukhin
ilblackdragon@gmail.com
1,457,663,107,000
c02b43e9b84dd57aeecb83a1eed9a4d1c3b18c1b
Replace pip install scipy with apt-get install python-scipy
Illia Polosukhin
ilblackdragon@gmail.com
1,457,669,152,000
7f4dc909c698cd1f9ea75c53bd46997edd0e3b68
Added guards around pandas requiring tests to skip them
Illia Polosukhin
ilblackdragon@gmail.com
1,457,670,403,000
60943fe083e4865f6a007e2c13e195b00e7c9c16
Update README.md Scores obtained on running label_image does not display normalized value.
Sarath Shekkizhar
0.omailbin@gmail.com
1,457,680,182,000
21002277cee0630dc104c6a6a0738fed918711a1
Merge pull request #1460 from shekkizh/patch-1 Update README.md
Pete Warden
pete@petewarden.com
1,457,714,837,000
64dd5b58d52d37697d5beb68e2177b966108e0a7
Add SparseTensor support to tf.batch and friends. Change: 116914274
Eugene Brevdo
ebrevdo@gmail.com
1,457,651,800,000
56f1d64998744ad655fe5c428658a13be35b865e
Fix dependencies bugs Change: 116925769
Eugene Brevdo
ebrevdo@gmail.com
1,457,659,110,000
14d82dd1b644b69761a6ec67cb76c878e2c2e058
Fixing threading issue with _maybe_load_sdca_ops(). Change: 116927408
A. Unique TensorFlower
nobody@tensorflow.org
1,457,660,630,000
c6e52ead31ccb45f66dccdb0c6399a63967fb0a1
Making SDCA updates transactional. Change: 116930935
A. Unique TensorFlower
nobody@tensorflow.org
1,457,663,721,000
625d62abef3d64ab9025cf248719b76aea3b7fc9
Guarding against 0-sized dense_weights (until we support arbitrary dimensional dense_weights). Change: 116936547
A. Unique TensorFlower
nobody@tensorflow.org
1,457,670,572,000
fc53648e0c0f5110bfab75a02b9dc75260b913d3
Use faster floating point rounding and floor primitives in a couple of image processing kernels. Specifically 1) round(x) -> lrintf(x) 2) int i = static_cast<int>(floor(x)) -> int i = x; for x >=0. 1) lrintf() is much faster than round() as it compiles down to a single instruction. Thanks to sesse@ for the tip. 2) C...
A. Unique TensorFlower
nobody@tensorflow.org
1,457,716,102,000
0e1a324c7bb73d80e0e2d069396d61115e0d9096
Improved the performance of the slicing operation on GPU Change: 116979767
Benoit Steiner
benoit.steiner.goog@gmail.com
1,457,717,339,000
f6f77accbb9c978b5d3562922425bffd01690f79
Implements "bool strictness" for `tf.Tensor`. Using a `tf.Tensor` as a Python boolean value is ambiguous: in library code it typically means "is this tensor not None?", whereas in user code it can mean "does this tensor evaluate to `True`" (which does not work and leads to buggy code). This change adds a warning when...
Derek Murray
mrry@google.com
1,457,720,454,000
a2b966692ffec38d5e6f1e35cf21ef9ab0ea01fc
Fixed a bug in the mean reduction code for GPUs Change: 116988587
Benoit Steiner
benoit.steiner.goog@gmail.com
1,457,722,777,000
e774fa829a1d74ab2d410cefc62cbd47e20b0317
Changes to make custom user ops not statically include tensorflow/core:framework. - Made a header only cc_library target for tensorflow/core:framework - Added a tf_custom_op_library macro to cut down the boiler plate - Added ability to compile custom user ops using CUDA Fixes #1378, #1419 Change: 116991629
Manjunath Kudlur
keveman@gmail.com
1,457,724,621,000
0249729a26b2cd3bdbbc5880f8d634a56860b0fd
Merge commit for internal changes
Manjunath Kudlur
keveman@gmail.com
1,457,730,970,000
f94a73c854b29aede9126354c09d73617a4bcea3
fixes #1293 - diag and diag_part gradients
Alan Wu
alan.c.wu@gmail.com
1,457,577,389,000
9e256f2f4eaf736bbdaef307ba87b066bc472fa5
Fix typos
Johnny Lim
izeye@naver.com
1,457,752,956,000
8e15958400dd9bf6cda93fb93373f13eca5962f2
pip test-on-install changes for skflow 1) Expanding the search scope for Python unit tests in order to include the new tests in python/skflow/tests 2) By default, installing scipy and sklearn in virtualenv for the aforementioned Python unit tests
Shanqing Cai
cais@google.com
1,457,753,762,000
848497bcef8a56b6d207da1ebea46812c2c3b2ca
Pull up a global statement in sdca_ops. The code worked before, but threw warnings about using a variable before the global call. This just hoists the global declaration to get rid of the warning. :) Change: 117002039
Craig Citro
craigcitro@google.com
1,457,731,497,000
464addfc500c36601b1c459937925847ba45f47a
Rollforward of "Add MirrorPad op. This op is a variety of Pad op implementing reflect and symmetric modes of Numpy pad." Change: 117005618
A. Unique TensorFlower
nobody@tensorflow.org
1,457,733,751,000
e2ffd3ada962aaf9cc30377f41f349fbf0e4c534
Warn (and fail) about implicit size limit for unique op (< 2^31 elements). Added a TODO to make this polymorphic, because fixing it requires a change to the type signature to be able to return an int64 for indices and counts. Change: 117005788
David G. Andersen
dga@google.com
1,457,733,833,000
eb5602546b426610814e6cf33396e935921daca6
Add optional cc file whitelisting for operator kernel registration. This allows unused operators to be stripped out of the binary to save space, even if they're in the srcs list Change: 117008234
Andrew Harp
andrew.harp@gmail.com
1,457,735,282,000
dbc57b9068da9b823dbcc3101e599da6f65f658a
Add gradients and tests for TopKOp. Change: 117009023
A. Unique TensorFlower
nobody@tensorflow.org
1,457,735,742,000
65c9124086240616747e26e0aa5ef3412a3be55d
Release a new TensorBoard version Change: 117011137
Dan Mané
danmane@gmail.com
1,457,737,053,000
9917b53a41345cf30954a5a4b17985412f1acfde
Clean up some unused code to fix warnings in the Mac build. Change: 117011572
Josh Levenberg
josh11b@tensorflow.org
1,457,737,345,000
c5fc107300f7c89befcedf605622b571dd8b9f71
Use the untemplated version of OpKernelContext::op_device_context() so we don't need to #include tensorflow/core/common_runtime/gpu_device_context.h. Change: 117019921
Josh Levenberg
josh11b@tensorflow.org
1,457,742,896,000
9849c30bb441d8bc07b5a5bb2efe0640ce46c97c
Improve SimplePlacer error message when a device does not exist. This change distinguishes between the case where a device in the NodeDef refers to an unknown device, and when it refers to a device that exists but doesn't support a particular node. Change: 117022320
Derek Murray
mrry@google.com
1,457,744,640,000
08332fc3e535c4c31c1969349b8b02fd6486cbe2
Cleaning up int64 implicit cast warnings from fifo queue. Change: 117023195
David G. Andersen
dga@google.com
1,457,745,283,000
90b1700d0df12cd03a4bfb75743bcf60a3c90255
Allow variable_op_scope to accept reuse as parameter. Change: 117023761
A. Unique TensorFlower
nobody@tensorflow.org
1,457,745,798,000
3b55e1f4f4be8fd4a6a5084edf9daf01e0990c3c
Change safe_strto32 and safe_strto64 to accept StringPiece. Updates callers to pass the StringPiece values. Change: 117027762
A. Unique TensorFlower
nobody@tensorflow.org
1,457,749,765,000
f916ae47698c19c0fb3116e471abc39e9b788279
Allows Variables to be used in sparse_tensor_dense_matmul(). Change: 117028597
Zongheng Yang
zongheng.y@gmail.com
1,457,750,938,000
ab6ffc92992f12456d2378f872be17f0ed274083
TensorFlow: allow growth in the GPU BFC allocator. This allows an option to start the BFC allocator small and grow it over time as needed. This can lead to increased fragmentation, but the benefit is that only as much memory as "needed" is reserved. This option defaults to off, but can be turned on by passing an opt...
Vijay Vasudevan
vrv@google.com
1,457,756,966,000
b7d3cf66f3d1ad9d79f9985c2c2c36dd5580d6ae
Merge remote-tracking branch 'caisq/help-1445' into skflow
Illia Polosukhin
ilblackdragon@gmail.com
1,457,762,552,000
e95d24f68f9c22ff8d42098d8094cfa6f4560885
Per @caisq comments moved pip install into Docker files
Illia Polosukhin
ilblackdragon@gmail.com
1,457,763,068,000
b1745b8513286c7a124260fa915f7a9908aca0ba
Fix defaults for early_stopping and ignoring it when it's none in the BaseMonitor
Illia Polosukhin
ilblackdragon@gmail.com
1,457,763,417,000
8aa7daf9c4efdfa6eab1170a760a80b59a067b27
Fix dockerfile syntax
Illia Polosukhin
ilblackdragon@gmail.com
1,457,765,113,000
7c4587121968c6b697a25dc8d05dd7e790c15b51
TensorFlow: Make copy assignment operator return itself in BFC Allocator. Change: 117036072
Vijay Vasudevan
vrv@google.com
1,457,764,799,000
30b52579f6d66071ac7cdc7179e2c4aae3c9cb88
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,457,767,626,000
3a955eb32d919bde8580ff53fb4e753a9579b8f5
Merge pull request #1469 from izeye/polish-20160312 Fix typos
Vijay Vasudevan
vrv@google.com
1,457,811,818,000
3b49611933c109b6f589e19ab5827aa586587318
pip test-on-install changes for skflow 1) Expanding the search scope for Python unit tests in order to include the new tests in python/skflow/tests 2) By default, installing scipy and sklearn in virtualenv for the aforementioned Python unit tests 3) Enabling caching of virtualenv directory to speed up pip test-on-inst...
Shanqing Cai
cais@google.com
1,457,753,762,000
bc601df32478d22466a28cbbc2f666724eb60b4e
Add space before filename in the error message of maybe_download().
Abhinav Upadhyay
er.abhinav.upadhyay@gmail.com
1,457,889,381,000
2f97f6aadce66027b04e8de1e1ba744b125dcbb9
Avoiding hard-coded bash path Changed '#!/bin/bash' to '#!/usr/bin/env bash'. There do not exist bash in most Unix system, like bsd family.
Iblis Lin
e196819@hotmail.com
1,457,845,769,000
ba892f05a32125962ec6cc829641220afab63cd1
Merge pull request #1484 from abhinav-upadhyay/assignment_fix Add space before filename in the error message of maybe_download().
Vijay Vasudevan
vrv@google.com
1,457,894,021,000
9c2f224adaa51bc348492ce265e54382f37933a6
Merge pull request #1472 from caisq/help-1445 pip test-on-install changes for skflow
Jan Prach
jendap@gmail.com
1,457,895,992,000
e6afc2f98a6c4a7ba760d47d029810ebb1eaedf3
Merge pull request #1478 from iblis17/no-hard-coded-bash Hard-coded bash path
Jan Prach
jendap@gmail.com
1,457,896,075,000
e8e0a83270908eea2d2367bc6699bb9f8ccd0ee0
Added GPU implementation for resize nearest neighbor grad.
panmari
smmuzi@gmail.com
1,452,786,033,000
968078a8532aadcec062e0c7fca239dd9698a88b
make xrange python 2 and 3 compatible in sdca_ops.py
Alan Wu
alan.c.wu@gmail.com
1,457,902,139,000
ca908c4bd1feaddb9a10d365a52d350f96b26131
Fix crash in BFC allocator rendering. Without this, gpf_bfc_allocator_test crashes when run with --v=4 because the region rendering code overflows the rendering buffer. Change: 117057531
A. Unique TensorFlower
nobody@tensorflow.org
1,457,816,911,000
d8d3786a94b82f72a8e8e9e5c27064b2ff58cf27
Reduce logging. Change: 117087996
Yuan Yu
yuanbyu@google.com
1,457,896,067,000
1cf9e29ec5f849eecb8d9700198585660a0814a1
Move notMNIST to Google Cloud Storage, update notebook. Fixes #1475. Change: 117088197
Yaroslav Bulatov
yaroslavvb@google.com
1,457,896,634,000
d2a06c2df73ed09d9eff60efb559fd38bc47e2d3
TensorFlow: fix compilation error on RPi Change: 117090447
Vijay Vasudevan
vrv@google.com
1,457,902,367,000
750832d12f72fd743e0cb2a8f56d5d7b42c3733d
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,457,905,459,000
ef16d7c929308a45c9c1ade0ff6874982cde882a
Fix WORKSPACE link, recommend submodule update
Jack Zhang
moonboots08@gmail.com
1,457,927,444,000
5c61053b2a2037b2e955041c1df149a90620e80d
Merge pull request #1490 from moonboots/master Fix broken android example README link
Vijay Vasudevan
vrv@google.com
1,457,928,854,000
42485650f65a275f46df1fc079e209a909da78e0
Merge pull request #1487 from chemelnucfin/range make xrange python 2 and 3 compatible in sdca_ops.py
Vijay Vasudevan
vrv@google.com
1,457,929,942,000
7ec8e0d5ecc7949da0c49613ec4b6116b2f47d3b
fix crash for empty dev_set bucket
Yuwen Yan
ybbaigo@gmail.com
1,457,970,194,000
7ea9843501d8938b7d58c5a95eacc3158b5784ec
Optimize `tf.nn.embedding_lookup()` and `tf.gather()` when shapes are known. This avoids cross-device transfers of shape metadata, which is often statically known at graph construction time. As a result, the load on the parameter servers is reduced. Change: 117135698
Derek Murray
mrry@google.com
1,457,969,682,000
bb06188d965cd4ad37d83c94c65d4db552dabc7b
Update Docker image to point to new data source. Change: 117140354
Vincent Vanhoucke
vanhoucke@google.com
1,457,972,947,000
e4add493f1020c0eb986aba21c266d9e6e6f4182
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,457,974,324,000
6778e5d7d02bc48587be61f1b3e0b4cd92b62888
Split tests for each loss type into separate test classes. Change: 117142191
A. Unique TensorFlower
nobody@tensorflow.org
1,457,974,090,000
d3f6604c622c7a6cf38bd266ff6626bfc960299c
Including cuda_kernel_helper.h, deleting some (now) duplicated code.
panmari
smmuzi@gmail.com
1,457,901,084,000
bc27f89b872153d0517a7881da13124f56b8c2ab
De-flakify testBatchNormArbitraryShapes Change: 117146617
Vincent Vanhoucke
vanhoucke@google.com
1,457,976,661,000
3d965a2df83764362f37d9c33e1b9e35d65eb57f
Fix minimap due to how the new polymer version handles shadow dom styles. Change: 117147170
Dan Smilkov
dsmilkov@gmail.com
1,457,976,948,000
419a66709c2c9eab5a23ba1ba9e987b5e81b005d
Eliminate now-unused pin_to_cpu code. (Legacy from before device("/cpu:0") did the right thing when nested.) Change: 117147187
David G. Andersen
dga@google.com
1,457,976,953,000
80344f3f7d22b39b5b74d93dabe7ba66cc32b214
Remove race conditions from TensorShapeUtils::MakeShape Also move some functions from header to C++. Change: 117148807
Geoffrey Irving
geoffreyi@google.com
1,457,977,856,000
b54ad5788b7cec515d81e9f7e3c9811fd8181235
Expose RunOptions and RunOutputs to Python's Run(). The Run() call that takes these protos previously only worked in C++. Example usage in Python (see also new unit test in session_test.py): sess.run(constant_op.constant(1.0), options=run_options, run_outputs=run_outputs) A thin T...
Zongheng Yang
zongheng.y@gmail.com
1,457,980,054,000
0bcaf2781eace3731c0ac07dea82920b07b6c2a4
Another 1 in 1000 numerical failure. Change: 117155975
Vincent Vanhoucke
vanhoucke@google.com
1,457,981,768,000
79453c19714cd8bc19b445fb7e29478e868ae4e2
Add bug id for bazel alwayslink issue. Change: 117162965
Josh Levenberg
josh11b@tensorflow.org
1,457,986,721,000
defc7870810c9503ebad3626f9f9a411a53597d9
Ensure the table size is returned only when the table is initialized. Change: 117166443
Yutaka Leon
yutaka.leon@gmail.com
1,457,988,673,000
c552a85b3bae3ce4c29c75aef8464346bc6b3088
correct commands pip -> pip3 Change: 117167194
Dan Smilkov
dsmilkov@gmail.com
1,457,988,954,000
ca250b2f64430c361c0699fc03aff9c91ddd3e55
Remove race conditions from TensorShapeUtils::MakeShape Also move some functions from header to C++. Change: 117167529
Geoffrey Irving
geoffreyi@google.com
1,457,989,086,000
86ddfc7a007d7e26f7c00df488969e302f5cd7a1
Moved session_manager and supervisor to third_party. Change: 117170777
Sherry Moore
sherrym@google.com
1,457,990,459,000
49792ef0a608da81807647ed4c23ab954398bd43
Add the test runner run_and_gather_logs to tools/test * includes a unit test that runs a benchmark and collects the results. * CC tests in tf/core/kernels now set linkstatic so GPU-based CC tests can be run by this test runner. * refactored the py test declarations in tensorflow.bzl to allow users to set main and args...
Eugene Brevdo
ebrevdo@gmail.com
1,457,990,541,000
cbba0465d73b06e5be123415cdbd12e61b6b1343
Handle the case when the same custom op library is loaded twice via load_op_library. * In C++, check if the library is loaded, and if so, return an error status. * In Python, memoize library filename to module mapping. If library already loaded, return the memoized module. Fixes #1378 Change: 117171149
Manjunath Kudlur
keveman@gmail.com
1,457,990,639,000
a465e286ccc9c83cb44f3153e91a8ae937904c49
Fix that TF_Buffer declaration erroneously makes use of C++ syntax. Change: 117171668
Zongheng Yang
zongheng.y@gmail.com
1,457,990,887,000
b44a68232087284eccd55c6031e7fc3ebe5663ab
Merge commit for internal changes
Manjunath Kudlur
keveman@gmail.com
1,457,992,685,000
0df350ce40590bab0d41d5f3a4222c1a08ed535e
Merge branch 'master' into translate_fix
Yuwen Yan
ybbaigo@gmail.com
1,457,996,108,000
c801f3bf00073ad49bb92b36e3050399cce092a4
Merge remote-tracking branch 'origin/master' into skflow
Illia Polosukhin
ilblackdragon@gmail.com
1,458,000,156,000
0ec0a9bc0437f069ea0b08c46f1734ab837d2bf6
Unroll partial merge from #1472
Illia Polosukhin
ilblackdragon@gmail.com
1,458,000,342,000
ba4c2112944f0505b309da538a6470a5f4f0d3a4
Disable run_and_gather_logs_test The test is broken until proper dependencies are declared in workspace.bzl.
Martin Wicke
martin.wicke@gmail.com
1,458,008,971,000
b9e9c9a12dbd2946d44c76abdd603eedf6eedb9a
Merge pull request #1504 from tensorflow/martinwicke-patch-1 Disable run_and_gather_logs_test
Martin Wicke
martin.wicke@gmail.com
1,458,010,634,000
bc8d5fe435e7489f904c961911727ecbc9f49d8f
quick python3 fix
Alan Wu
alan.c.wu@gmail.com
1,458,015,581,000
2f96fae6069db06ab7b2999d97ec97c68f385cfe
Merge remote-tracking branch 'origin/master' into skflow
Illia Polosukhin
ilblackdragon@gmail.com
1,458,015,812,000
cf95926468832b590d7713f5a15648be25593e5e
Merge branch 'master' into translate_fix
Yuwen Yan
ybbaigo@gmail.com
1,458,017,440,000
69a5419a61a37a2674b4e6573e9a51828c5d1e9f
Merge pull request #1507 from chemelnucfin/quickfix quick python3 fix
Martin Wicke
martin.wicke@gmail.com
1,458,017,998,000
14b64913a90229ab5edd346a1a92c8f57c8c4d63
Fix python3 b
Martin Wicke
martin.wicke@gmail.com
1,458,019,822,000
ad0be489019f97e1b5cf25c51a502d80456672d2
Added skflow to contrib_py build rule. Removed cyclical dependency
Illia Polosukhin
ilblackdragon@gmail.com
1,458,023,028,000
96efe662fe31ab4ea739702e5781b1edc2e2aecd
Changed to absolute paths in __init__ for skflow
Illia Polosukhin
ilblackdragon@gmail.com
1,458,024,169,000
318814d6e8ba01d4f20385c6773bbb41ee60bea3
Merge pull request #1509 from tensorflow/martinwicke-patch-2 Fix python3 b
Martin Wicke
martin.wicke@gmail.com
1,458,025,093,000