hash
stringlengths
40
40
msg
stringlengths
1
131k
author
stringlengths
1
33
email
stringlengths
0
57
date
int64
1,447B
1,698B
17d3da1c24fe1542ff4ca333403f3b4447039cac
Merge pull request #1219 from NHDaly/patch-1.
Vijay Vasudevan
vrv@google.com
1,456,163,925,000
82d86d9c002939f867b8dce309f91b1b57309493
Update RELEASE.md for 0.7.1
Vijay Vasudevan
vrv@google.com
1,456,162,871,000
657b5f398901d0a36a5cac15dd62d27752a69b28
Merge pull request #1238 from vrv/r0.7 Update RELEASE.md for 0.7.1
Martin Wicke
martin.wicke@gmail.com
1,456,164,533,000
e7c34a5b19f8fd53e7e4b12708cff8e0c7a4770b
Add cache_device to Variable construction and variable_scope. This is necessary for, e.g., RNN where one wants to cache Variables locally even when they are accessed through a conditional like cond. Without the local caching, each cond creates a Switch that bypasses the current Variable copy deduplication code and fo...
Eugene Brevdo
ebrevdo@gmail.com
1,456,017,957,000
f0c52770d61ee1c925477de2b0e2d72d1c5dbb6d
Add a few more special GPU kernels for int32 and bool. This improves the performance of dynamic RNN into the reasonable range when comparing with cond-based static unrolling. Before: Graph Creation: Static Unroll vs. Dynamic Unroll LSTM max_t dt(static) dt(dynamic) dt(dynamic)/dt(static) 1 0.973783 1.598944...
Yuan Yu
yuanbyu@google.com
1,456,091,399,000
1d18505293359a2506670362592a027bc1a9ace0
Replace info-card non-printable whitespace with <wbr> Change: 115220444
A. Unique TensorFlower
nobody@tensorflow.org
1,456,153,234,000
41b894ae93019f1ae69065033c7c1210ddb605ca
Adds functionality to the platform library. The testing::SrcDir() function supports data dependencies in cc_test targets. Change: 115239392
Derek Murray
mrry@google.com
1,456,163,873,000
2edddcf1123f050f9ddf27b03a2226aec42f12f1
Adds implementations of PosixEnv::SchedClosure[After](). The current implementations are very simplistic: they spawn a thread in response to each call. This is needed to deal with the fact that some users of SchedClosure rely on the ability to spawn blocking closures, and we lack an unbounded threadpool. TODO(mrry): ...
Derek Murray
mrry@google.com
1,456,163,978,000
0e9240b72613aa084728dcba55f7fae3591e86ad
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,456,165,757,000
8dbe7bbb99a76666cc58e38532086b339c9c2100
Merge branch 'master' into r0.7-merge
Vijay Vasudevan
vrv@google.com
1,456,168,375,000
05c58647731480152b295c0a3f316f21a0641c14
Merge pull request #1239 from vrv/r0.7-merge Merge branch from 0.7.1 into master
Vijay Vasudevan
vrv@google.com
1,456,173,195,000
18617f9ef1f558f872c464aaa5819afc2213addb
Add dummy Variable.__iter__ to avoid endless hang when accidently treating a var as a list. Change: 115243053
A. Unique TensorFlower
nobody@tensorflow.org
1,456,165,595,000
12b36c655359319c1ad31907d407bba3c594fee7
Add operator overload for !=. Change: 115243253
A. Unique TensorFlower
nobody@tensorflow.org
1,456,165,683,000
24d64cf2d52d5fca8c78fc2ce065737fceb0b461
Add shape function for RandomCrop. The absence of shape function makes import_graph_def() fail when RandomCrop exists in GraphDef. Change: 115243268
A. Unique TensorFlower
nobody@tensorflow.org
1,456,165,691,000
26903ee7b6c80d2f6029ede38b49e4a044e6b9b9
Fix division by 0 bug in EventAccumulator. Change: 115249194
A. Unique TensorFlower
nobody@tensorflow.org
1,456,168,522,000
8fd0027e83b5da279f0ddf1aa6f100d9f76eeb62
Added an "mark_as_used" option to unique_name() to control whether the call changes state. If it's set to False, the new name that will be used will be returned without actually marking the name as having been used. Change: 115249981
Sherry Moore
sherrym@google.com
1,456,168,894,000
2c404fe881c4e24ebba616a2fee2a77ac0eed234
Add IsAligned method to Tensor. This is useful when you produce tensors using Split, but don't want to copy them unnecessarily if you need them aligned. Without this, you need to pessimistically assume that they are all unaligned and copy all of them. Change: 115251844
A. Unique TensorFlower
nobody@tensorflow.org
1,456,169,823,000
25621e49d8dcc5edd14a3bcc2d56b8c8cf1be9e9
Added support for half floats to eigen, which is the first step to support half floats in TensorFlow. The code was tested on Tegra x1. Change: 115253733
Benoit Steiner
benoit.steiner.goog@gmail.com
1,456,170,833,000
87cabf37f326838c69d2cf196c123851f32f0ffd
Change the API for node-radar so instead of having many callbacks, it returns metadata at a single point of control. Change: 115255052
A. Unique TensorFlower
nobody@tensorflow.org
1,456,171,490,000
0bc73c25a56df622304091b2d2434e0d2b5e4ca7
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,456,175,883,000
5aa3965ba44cd9a0da68022d00c1a5de3425c875
Fix the license in tensor_format.cc. Change: 115261957
Xiaoqiang Zheng
zhengxq@google.com
1,456,175,502,000
8f8f7f11b9e49ed8fb297ec5b5cf5be30f93359d
Added the ability to track allocation sizes in the tracking allocator itself if the underlying allocator doesn't already do it. Change: 115263741
Benoit Steiner
benoit.steiner.goog@gmail.com
1,456,176,595,000
a3841eeeeb1d0d11df732c560c50f5e81f768cd9
Upgraded to the latest version of eigen, which adds a missing #include Change: 115268843
Benoit Steiner
benoit.steiner.goog@gmail.com
1,456,179,204,000
ec140b413f62fe6e13d436a4de85228bed26055a
Correct device mis-placement for switch with apply_gradients. Change: 115269320
Lukasz Kaiser
lukaszkaiser@gmail.com
1,456,179,458,000
225de5e336d4d241ce02652eb961cdb6e0eec847
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,456,179,953,000
6c9c4d7118d966648924b6e261b05f2fd139c4fe
Make the TensorFlow Serving tutorial layout consistent with the other tutorials on tensorflow.org. Change: 115270889
Jarek Wilkiewicz
jarekw@google.com
1,456,180,250,000
4bad5ba5b1bcaf5c727facb7ac1b6f10e5aa5f30
Get rid of attempt to adaptively use Cholesky for inverting SPD matrices in matrix_inverse. Cholesky is not enough faster than PartialPivLU in Eigen to be worth the extra pass over the data to check for symmetry, even for relatively large matrices. For (n=2k) SPD matrices Cholesky does give ~5% speedup over LU, but for...
A. Unique TensorFlower
nobody@tensorflow.org
1,456,183,382,000
52fce2481d21800a7a6483286e83315be4645bba
Upgraded Eigen to fix a compilation error triggered by xcode Change: 115280348
Benoit Steiner
benoit.steiner.goog@gmail.com
1,456,185,265,000
aea1bf524b5b5026b2bcc5439c91d886afa827d4
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,456,185,496,000
993c77de77ec311f4f68f756fdccb5e27e2b6345
Avoid copying invalid large constant data. Tensorflow cannot create an operation with a NodeDef larger than 2GB. (Python Unit Test Case: ConstantTest.testTooLargeConstant) However, before throwing a meaningful exception, it consumes lots of memory to copy the invalid input data. This causes `std::bad_alloc` errors som...
Dongjoon Hyun
dongjoon@apache.org
1,455,923,944,000
639d48f1f225e5dba865a5eeef513dbe88d6470e
Merge pull request #1207 from dongjoon-hyun/avoid_copying_invalid_large_constant_data.
Manjunath Kudlur
keveman@gmail.com
1,456,205,310,000
e9a6432014d5f1e64819baa96dfb6e768d326b7d
Fixed learning_rate in trainer
liyongsea
lijia0765@gmail.com
1,456,248,082,000
5bd9073799436daa9275a2990e5df58a75a1aa78
Add more control over shapes when calculating graph metrics. Change: 115284554
Pete Warden
pete@petewarden.com
1,456,187,636,000
93866d1407f31011d307bf3868ad72516419fea0
Updated image retraining documentation with usage example Change: 115288487
Pete Warden
pete@petewarden.com
1,456,190,057,000
3972b66fa5e3bd89e50fc01b43c16b81b52e9319
Add license header and future imports to sparse_tensor_dense_matmul_grad_test. This should fix Python 3 compatibility for this test. Change: 115339521
Derek Murray
mrry@google.com
1,456,242,732,000
05d950331e23bc57d7c4739ad98259f472aa0664
Adds a C++ test utility that makes it possible to run multi-process tests. This will be necessary for tests the distributed runtime (issue #23). Change: 115339579
Derek Murray
mrry@google.com
1,456,242,779,000
cc36921c64c88857757d80bd7d562042cde50e1f
Adding tensor shapes to the graph visualizer and to the info card. Also: - rename sceneBehavior -> sceneElement to make it clearer it is a polymer element. - improve the info card by showing the actual op node in the successors/predecessors list when the metaedge only contains one base edge (one tensor). Change: 11533...
Dan Smilkov
dsmilkov@gmail.com
1,456,242,971,000
35fa8c4ef0a60f854e40491518153260dde0b8da
Adds a C++ test utility for picking an unused port. The PickUnusedPortOrDie implementation is based on a simplified version of `grpc_pick_unused_port_or_die()` in gRPC. This utility will be necessary for tests of the distributed runtime (issue #23). Change: 115345502
Derek Murray
mrry@google.com
1,456,246,729,000
082263066a23afed631fc04462784773e68d3d71
Reimplement StrippedOpListForGraph in C++ Change: 115347996
Geoffrey Irving
geoffreyi@google.com
1,456,247,839,000
b1496f68f34d77bd958a27df4d51cbe676f2d370
Disable shape inference for accumulator when the shape is not fully known statically. Change: 115351830
Yuan Yu
yuanbyu@google.com
1,456,250,016,000
6643b804b561a3e0776bddee0c0dd6e55e90a35a
Update graph visualizer tutorial to include new interactions. Change: 115354844
James Wexler
james_wexler@yahoo.com
1,456,251,495,000
2b9f60917a255137d9e97c8e4e767718b07d9152
Fix build rules for C++ tests in the OSS build. Change: 115358623
Derek Murray
mrry@google.com
1,456,253,217,000
b8d35c6aa7996efdb0f96f88be0941ae98308598
Merge pull request #109 from liyongsea/fix_learning_rate_name_in_trainer Fixed learning_rate in trainer
Yuan (Terry) Tang
terrytangyuan@users.noreply.github.com
1,456,255,685,000
803f9a4badfe0203550fdaebd651f23c9fc1b216
Memory swapping between GPU and CPU for stack. Changed stack push and pop ops to async. For gradient computation for loops, stacks are used to store the tensors that are computed in the forward but needed in backprop. This CL enables very long sequence training by swapping the stack tensors from GPU to CPU. Change: 11...
Yuan Yu
yuanbyu@google.com
1,456,253,851,000
986762835b7484ded2893ebee967c77857d6b970
Adding tests for TensorFlow tutorials The tutorial Python files are copied to a separate directory and run against Python installation on the system. The script performs basic quality checks, including timeout, accuracy / loss thresholding, and checks for generated checkpoint files and summary files. Change: 115362939
A. Unique TensorFlower
nobody@tensorflow.org
1,456,255,259,000
91e1180fac8defe3cc5ea91509d3658d5a08f61a
TensorFlow gflags: add DEFINE_bool as an alias to DEFINE_boolean Change: 115364038
Vijay Vasudevan
vrv@google.com
1,456,255,789,000
cdce29ed7959149abd91258ecf4308e5114ea72b
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,456,255,933,000
fb0a3f20d9e20853bb1fefebac6cb0a1e0f637a6
Swap TensorBoard local build from "tsd" to "typings". Reason: tsd is deprecated (https://github.com/DefinitelyTyped/tsd/issues/269) and typings is the new standard. Also, tsd was behaving badly - running `tsd install` on a clean client was causing it to incorrectly depend on typing files from node_modules, which result...
A. Unique TensorFlower
nobody@tensorflow.org
1,456,259,231,000
c9b6b3509f15da16028eb865ddff0331da9d8aaa
Fix tslint errors in TensorBoard Change: 115370821
A. Unique TensorFlower
nobody@tensorflow.org
1,456,259,378,000
632a7fdcb98cdccb29bb5a3dae75e73fba6daf67
Add tensorboard graph visualization link to graph visualization how-to Change: 115371065
James Wexler
james_wexler@yahoo.com
1,456,259,521,000
4aa77788f5fc00c567ee6cd8389508fa55b4bf37
Merge commit for internal changes
Vijay Vasudevan
vrv@google.com
1,456,261,470,000
e7c6ffde2e212c2b6998fbb162ba60c5f5026573
Update protobuf submodule to bring in gRPC support
Manjunath Kudlur
keveman@gmail.com
1,456,263,105,000
17f22bb5bd65b40c912039ee7360bb6c04202e5c
update_version.sh: automatic version update in source Usage example: update_version.sh 0.8.0
Shanqing Cai
cais@google.com
1,456,247,346,000
72b5387d16f6e2f292440fdb486cb7838f2c2927
Merge pull request #1255 from caisq/automate-ver-1.
Martin Wicke
wicke@google.com
1,456,263,259,000
0a25618a783e43e0f11f3c5de628b81e939d17fa
Add nightly binaries to README.md
Martin Wicke
martin.wicke@gmail.com
1,456,263,357,000
634c37e65928c2556774d4273060b637dbcb9bc8
Merge pull request #1218 from tensorflow/add-nightlies.
Martin Wicke
wicke@google.com
1,456,263,400,000
e7d00e3bed71af45614a7d733e081804b386cb8d
Merge commit for internal changes
Manjunath Kudlur
keveman@gmail.com
1,456,264,313,000
e59be7083e15d83de731e3c8e2a4889f5b34114f
Adding switch for tutorial tests to ci_parameterized_build.sh
Shanqing Cai
cais@google.com
1,456,265,475,000
bfd5f0bfaa13f397e80063ce59951033928b04b6
Merge pull request #1259 from caisq/tut-test-1.
Martin Wicke
wicke@google.com
1,456,265,485,000
83cfb3943c9893fc9850e840d14551e910fbb903
Add early stopping and reporting based on validation data
Dan Becker
dansbecker@gmail.com
1,455,597,518,000
623fac64f5b33bb0c1f691eb7c5fc58798eeefb3
Minor refactor of monitoring classes. Add monitor arg to estimators. Related fix-ups
Dan Becker
dansbecker@gmail.com
1,455,926,964,000
f933c9956ee54dab78cae9ff83b11c2d09eebd8a
Move monitor argument from model specification to fit
Dan Becker
dansbecker@gmail.com
1,456,110,450,000
87a893157c8dfcd19cb53bea5dc154a9b9c6831e
Update early stopping test for consistency with new api
Dan Becker
dansbecker@gmail.com
1,456,111,947,000
f85dc7cba8d7b7f79b34e2d4f7f973363547d383
Update early stopping example
Dan Becker
dansbecker@gmail.com
1,456,168,311,000
9b488c5d150b50db57f2e8560b37f0c1e4f0c92d
Add monitor.BaseMonitor and arg pass through from val monitor
Dan Becker
dansbecker@gmail.com
1,456,278,091,000
1bd73a86c4285594ca31181f49ff7c56c60ce924
Update 3_regularization.ipynb Fixed a mistake in a variable name.
urimend
urimend@gmail.com
1,456,312,159,000
b0dc4b70c532d92ef86017253c5fb6861ab9f2c7
Merge pull request #1272 from urimend/patch-1.
Vincent Vanhoucke
vanhoucke@google.com
1,456,324,105,000
972fc42b93be51c8ebaa6c5c536a454085958958
Adding permission x to test_tutorials.sh
Shanqing Cai
cais@google.com
1,456,324,875,000
3159f6e8337d15232b98c2325fdd8dc4133366e8
Fix lint
Dan Becker
dansbecker@gmail.com
1,456,332,312,000
a28882f7fc50a1d59f0bd227ff5d2d7ddc42b451
Merge pull request #104 from dansbecker/validation-early-stopping Fixes #85 Add early stopping and reporting based on validation data
Illia Polosukhin
ilblackdragon@gmail.com
1,456,338,370,000
00ccdc58e11b8a6554c99ca113ee95787bcba677
Merge branch 'master' into patch-1
jmtatsch
tatsch@gmail.com
1,456,348,485,000
2408e359cc9e51bedf561fc62c0a40e337fa63b9
Give tensorflow/core/kernels/ its own BUILD file. Change: 115379524
Josh Levenberg
josh11b@tensorflow.org
1,456,264,402,000
185cff7f41600e26d5f9ecfb18bcaab09edf4bfe
Make core/framework/graph_def_util.h publicly accessible. Change: 115384748
Josh Levenberg
josh11b@tensorflow.org
1,456,266,924,000
94a992cfc3266244b81fe311b805fc1ae3f53f30
Add correct dependencies to sdca ops to fix build breakage. Change: 115408162
A. Unique TensorFlower
nobody@tensorflow.org
1,456,282,279,000
5f4ec004b89e63e6501e4e9b0aadbbb15ebc58b6
Tests to check linear_optimizer in tf.contrib. Change: 115419426
A. Unique TensorFlower
nobody@tensorflow.org
1,456,296,054,000
92383c8754179375ef1e91f270cd60a126cf77c4
Fix minor typo in documentation in training_util.py Change: 115462062
A. Unique TensorFlower
nobody@tensorflow.org
1,456,335,992,000
9d84271a2039918994e57c2f962d2ee656f01541
TensorFlow: Initial support in SimplePlacer for colocation groups, to be used to colocate based on attributes rather than either names of ops or devices (op names and devices aren't portable). A follow up change will add an ops.colocate_with() to Python that adds this attribute to nodes, and will be used to replace ca...
Vijay Vasudevan
vrv@google.com
1,456,336,790,000
8411effdee68041ebccee422095444a0753d344b
Add the OneHot op. Change: 115464229
Eugene Brevdo
ebrevdo@gmail.com
1,456,337,169,000
8804a486c74f96d5cc885b70f3950dff195e48fc
Store only what is needed in the node name to node map. Change: 115464489
Yuan Yu
yuanbyu@google.com
1,456,337,289,000
26078dfaf20f20a5c5d8a30503fbf387d7d30d53
Fix safety bug in gather and scatter Both gather and scatter now unconditionally validate indices in the inner loop, which prevents crashes if indices are changed asynchronously while the ops are running. For gather when validate_indices = true, the new code is within the noise of the old code speedwise or possibly s...
Geoffrey Irving
geoffreyi@google.com
1,456,338,604,000
14a237beb0f917103ae82bcff298fbda7eb6ed39
Update TensorBoard README.md. Describe how to load many runs. Change: 115467346
A. Unique TensorFlower
nobody@tensorflow.org
1,456,338,754,000
6b2c0012d1caa3026641f5d5e83e9ba7d203ab61
Eliminate unneded pylint disable Change: 115470945
David G. Andersen
dga@google.com
1,456,340,566,000
57df84c47eac324418b2bf11d33cb4b50ba274b2
Rename map in control_flow_ops to map_fn, to avoid name conflict with Python's native 'map' function. This also fixes the bug with control_flow_ops.case Change: 115472163
A. Unique TensorFlower
nobody@tensorflow.org
1,456,341,133,000
87a289103fce76e7224f2023332efe6f92ce6ee6
removing repeated hostcast lines Change: 115472914
A. Unique TensorFlower
nobody@tensorflow.org
1,456,341,509,000
4afef14f02682433f661b31ba57018d4ac6bdc18
Support leaving the offset (beta) parameter out in batch_normalization, in which case no offset will be added after normalization. Change: 115489328
A. Unique TensorFlower
nobody@tensorflow.org
1,456,350,839,000
746ccc842e84ce1f4198be8f0ead860d22071b96
Temporarily disable sdca_ops_test - it breaks the opensource build. Change: 115494526
Eugene Brevdo
ebrevdo@gmail.com
1,456,353,382,000
497606904be87f7a4078ed7ee0784afaa094b258
Fix build issue with safety fix to gather and scatter Change: 115495726
Geoffrey Irving
geoffreyi@google.com
1,456,354,003,000
2861cc1d239932a5dfd745c86f30acc8bb947b6d
Surface control_flow_ops.case to public. Update docs. Add unit tests. Change: 115496194
Eugene Brevdo
ebrevdo@gmail.com
1,456,354,249,000
223794ee783244977e865a3724904a39a35720ed
Avoid using initialization lists since the version of nvcc shipped with Tegra X1 crashes when attempting to compile them Change: 115500414
Benoit Steiner
benoit.steiner.goog@gmail.com
1,456,356,373,000
5c26a0aaebfc9fbf9c72f8abf671cc3ce8c84af7
Merge pull request #1275 from caisq/tut-test-2.
Vijay Vasudevan
vrv@google.com
1,456,377,858,000
8041c546bb7803dcb5f8c98877d2fe99c8966acb
Switch sdca_ops to use tf.load_library mechanism. Change: 115505008
A. Unique TensorFlower
nobody@tensorflow.org
1,456,358,944,000
2cc5ed87e3308f37cd1eaacc58992367f679e69f
Execute TODO to explain graph-consumer usage of RemoveNewDefaultAttrsFromGraphDef(). Change: 115506523
Josh Levenberg
josh11b@tensorflow.org
1,456,359,773,000
bce6216610d57f8f4b1e9e79836737df109c4e42
Switch nn.moments() to using a one-pass stable algorithm. Helps with: https://github.com/tensorflow/tensorflow/issues/917 Also fixes https://github.com/tensorflow/tensorflow/issues/1162 The main benefit is that the computation of the sufficient statistics is now decoupled of the aggregation of the moments, which means...
Vincent Vanhoucke
vanhoucke@google.com
1,456,360,074,000
ab286e09968c4a9c2032ab803d1b89be56f3e374
Remove no-longer-needed RequireDefaultOps(). Change: 115511794
Josh Levenberg
josh11b@tensorflow.org
1,456,362,942,000
9ba55d8a7558f6670fbf47eeebb90901b362d873
Remove no-longer-needed RequireDefaultOps(). Change: 115511835
Josh Levenberg
josh11b@tensorflow.org
1,456,362,972,000
c7cbf87735728e930f1b5f7afc53bd01f78a0d34
First version of CMake build files that compile and link the c++ example
Aurelien Geron
aurelien.geron@gmail.com
1,456,419,731,000
77da168dbc0d2324a8cf75a0933924eac2eab39f
Move all Tensorflow WORKSPACE rules to a skylark macro Change: 115515678
Kiril Gorovoy
kgorovoy@google.com
1,456,365,736,000
4ecd2a70dd750b20a61033fe08301745685bf288
Added unit test for max_to_keep being None. Change: 115516426
Sherry Moore
sherrym@google.com
1,456,366,247,000
fcfa866d673aaa08a63f2fd57b2ceafec823d9ec
Added TestReporter and test / benchmark reporting tools. These tools are meant to allow recording of benchmark & unit test structured output to pbtxt files in a directory only when the environment variable TEST_REPORT_FILE_PREFIX is set. For now, only saving of C++ microbenchmark output is supported. Change: 1155183...
Eugene Brevdo
ebrevdo@gmail.com
1,456,368,026,000
73d557cc88ee86834917088416e5ce8783d11798
Fix an error message in tf.sparse_to_dense to include the possibility that indices are invalid because they are out of bounds. Change: 115522264
A. Unique TensorFlower
nobody@tensorflow.org
1,456,372,881,000