hash stringlengths 40 40 | msg stringlengths 1 131k | author stringlengths 1 33 | email stringlengths 0 57 | date int64 1,447B 1,698B |
|---|---|---|---|---|
481c4ba7a8cc0f9924eed7578502657420609988 | Updated the distributed runtime docs to use `tf.GrpcServer`.
It is no longer necessary to build `grpc_tensorflow_server` to use the
distributed runtime. All dependencies are now included when you
`import tensorflow` in Python.
This change also adds a convenience method for creating an in-process
server that binds to ... | Derek Murray | mrry@google.com | 1,458,585,609,000 |
03993ff2aeab67694ec3e14fb6b02c50efd8b287 | Apply selective registration to op registration, to remove the OpDefBuilder
calls.
Move selective registration macros to their own file - this made it easier to
#include <string.h> before the definition. Renamed SHOULD_REGISTER_OP to
SHOULD_REGISTER_OP_KERNEL, and used SHOULD_REGISTER_OP for the new op.h
filtering.
C... | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,586,920,000 |
0c83a3ca4f18c46c962c307f4ace39496ce8a850 | Adding end to end tests in TensorBoard.
Firing custom "rendered" event for the event, image, histogram and graph dashboards as a way to know that everything has rendered properly.
The graph dashboard fires "rendered" after some custom d3 logic, while the other components fire it when attached, using async, as suggest... | Dan Smilkov | dsmilkov@gmail.com | 1,458,588,311,000 |
5c1bcf1cfb61cc9750228820151a52f1d6426dc4 | Improve input validation of SparseSegmentReduction ops. Also add
validation that all output is written (or an error is generated) for
all the SegmentReduction ops.
Change: 117748193 | Josh Levenberg | josh11b@tensorflow.org | 1,458,592,364,000 |
59e04a09f53b560335aa23c8d4bfbf242feff075 | Adding more types for __ANDROID_TYPES_FULL__ profile.
Change: 117755581 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,596,172,000 |
a59e6e027a82703d337d2d4626f4a16799d9cab1 | Add a scan operator, similar to Theano's scan.
Change: 117756292 | Yuan Yu | yuanbyu@google.com | 1,458,596,515,000 |
44864a9eb83db763257d88f0814c2f8230145b34 | In jpeg Uncompress, expose nwarn as an int64* to be consistent with the
underlying library.
Change: 117761679 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,599,468,000 |
2a78729b4b2abc0f098fe7f21d6bdb862d8da138 | In DecodePng, verify that the width, height, and total size of the image to be
decoded are within a reasonable bounds.
Change: 117761753 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,599,513,000 |
b49474d2794a86b4f4e27aefb0998e1c7734135b | Rename RunOutputs -> RunMetadata proto.
Reason: it can be confusing to have session.run(ops, run_outputs=...) as users might get confused that run_outputs contains the outputs of the ops that are passed.
Or in the C++ API it is even more confusing since we have Session.Run(..., std::vector<Tensor>* outputs, RunOutput... | Dan Smilkov | dsmilkov@gmail.com | 1,458,601,117,000 |
6047c6977dbc30f018b8b3ea0486ca907901dabb | In DecodePng operator's Convert8to16 function, force height*row_bytes
computations to use 64 bits, to avoid overflow.
Change: 117766779 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,602,381,000 |
cb8070b4b50b21297ddb00c250cf1da31613aea9 | When variable op scope fails, give some indication about where the variable was originally defined.
Change: 117766780 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,602,383,000 |
dde0b44ab2559c5896cf313a2f790094373e488f | Enable tf.sign for complex64.
Change: 117767157 | RJ Ryan | rryan@alum.mit.edu | 1,458,602,625,000 |
6697febf679775a33623a6cfb9cdcdd2f438b561 | Write zero values for GPU gather when the index is out of bounds.
Change: 117767826 | Eugene Brevdo | ebrevdo@gmail.com | 1,458,603,003,000 |
7e572539f7b00d35dacfd5a11807fd302238f6a5 | Change jpeg decoding to record the number of rows processed, instead of the
fraction processed. Also avoid one case of int overflow.
Change: 117774225 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,607,095,000 |
bc41791709a11125439601d9af7f32fcf293a026 | Mark a bunch of tests "small" to reduce timeouts
This should be everything except for except for python, tensorboard, cc,
contrib, and tools.
Change: 117774679 | Geoffrey Irving | geoffreyi@google.com | 1,458,607,448,000 |
aadf300794cec394ef894e5d94a2ba3d5a5fcb9f | Add a SparseAdd op that adds up two SparseTensors.
This CL provides a CPU kernel implementation.
Change: 117778151 | Zongheng Yang | zongheng.y@gmail.com | 1,458,610,167,000 |
a8955c165929514f5ef6268010cbe198e4a32164 | Add GPU support to Variable and Assign for all POD types
Previously only float and double variables were allowed on the GPU. Now
anything other than string works. The same goes for Assign. AssignAdd
and AssignSub have been left alone for now.
Change: 117786325 | Geoffrey Irving | geoffreyi@google.com | 1,458,618,887,000 |
a91af1bdb9cfb74a689f252ac443de508e657118 | BUGFIX: Correct the naming scheme used in weighted_moving_average.
Now, the signature is (..., name=None) and variable scope uses name and
default_name correctly, allowing the default name to be uniquified.
Change: 117787114 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,619,649,000 |
38cc74049a10031e11d652587784b806514d702c | remove test_timeout from oss ci tests | Jan Prach | jendap@google.com | 1,458,259,060,000 |
ef3409d0eb76e9f3ec0c181616ab91330f4805b1 | more shards for python/kernel_tests | Jan Prach | jendap@google.com | 1,458,639,612,000 |
5337a25ad15bf4e7fbf1a87fe1698fe5b16f15b2 | added the first tests, and package organization | Vu Pham | phvu225@gmail.com | 1,458,640,173,000 |
bf554a4f4a9f6f7344a21143786e3c54a846db0a | Merge branch 'master' into enhancement/grid-rnn | Vu Pham | phvu225@gmail.com | 1,458,640,205,000 |
ae618fd9b1c3e4c5d8e9accd2ac42b2e1e763e19 | Update tutorials index. | znah | moralex@google.com | 1,458,644,099,000 |
34b4ce808b3161f125357d412db3268ef48884c1 | Update nbviewer link in README.md | znah | moralex@google.com | 1,458,645,671,000 |
7c85129205cb484d7f861fe822c19d10c2eb375d | Fix a Tensor-as-boolean issue in cifar10 tutorial | Shanqing Cai | cais@google.com | 1,458,655,614,000 |
1fb2d383040fdc6c8e198080b8f088568fe3b338 | Fix docs for cross-entropy loss functions | Robert DiPietro | rdipietro@gmail.com | 1,458,663,963,000 |
ad1d98011ebcedbce07216e9036313ec5a19fb52 | Merge pull request #1536 from jendap/ci_build-without-test_timeout
test without long test_timeout | Martin Wicke | martin.wicke@gmail.com | 1,458,670,151,000 |
b7e004cfe78f12bbe40f1204c4774e7cfc7e62ef | Merge branch 'master' into fix-python3-1 | caisq | cais@google.com | 1,458,671,077,000 |
ec9391d672f1f6239390efe8eb6b41aa374e8d19 | Merge pull request #1579 from caisq/fix-python3-1
Fix a Tensor-as-boolean issue in cifar10 tutorial | Martin Wicke | martin.wicke@gmail.com | 1,458,676,689,000 |
44e95baf4545c0b7e4ea7351da44e67de698c6a8 | Merge pull request #1567 from znah/deepdream
Moved deepdream to tutorials. | Vijay Vasudevan | vrv@google.com | 1,458,677,553,000 |
733321f6159f55474f4d7822b5a1c174d30e696d | Use tf.group in rnn benchmark.
Change: 117796203 | Yuan Yu | yuanbyu@google.com | 1,458,630,146,000 |
9d03824d6740be0c043c431566c917ec0cf6cf3f | Add a half type to TensorFlow core, based on Eigen::half.
Note that this is only the type, not support for it in any ops,
so it is not useful for anything yet. In particular,
neither TF_CALL_REAL_NUMBER_TYPES nor TF_CALL_GPU_NUMBER_TYPES
list Eigen::half, so even though a lot of ops will end up
declaring support for t... | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,660,911,000 |
20a316fbbd51faa01e3e52c2e2e64d45c55d4df3 | Mark more tests as size = "small"
Change: 117825911 | Geoffrey Irving | geoffreyi@google.com | 1,458,661,252,000 |
d48632335f62890b29d57744cab118056d3bb85a | Expose filegroups from the android example.
Change: 117827602 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,662,503,000 |
1dffd48b7351dfef7ee6e1e53d6cb4390975ddf7 | Fix cifar tutorial for bool strictness.
Change: 117828297 | Martin Wicke | wicke@google.com | 1,458,662,883,000 |
e5f879c2a763f1f74e7078455b8d733f2d21702a | Override the default message size limits in TensorFlow gRPC.
This prevents errors when large tensor constants are added to the
graph, large values are fed to a step, and large values are fetched
from a step.
Thanks to @ms705 for uncovering this issue.
Change: 117830058 | Derek Murray | mrry@google.com | 1,458,664,112,000 |
c57ea3c70942f5fad0638dae60b5131d36463a67 | Adding a gradient to ComplexAbs.
Change: 117831771 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,665,170,000 |
0f64cbf389daa90e4c7335d29e6c9061bdce5fd5 | Remove size = "medium" default from tensorflow.bzl functions
We want explicit size arguments for all tests.
Change: 117832476 | Geoffrey Irving | geoffreyi@google.com | 1,458,665,607,000 |
723d6f8b5d1ce14c4a03ec3e2a430d86e6b51c74 | Changing get_binary_predictions_* methods in sdca_ops_test.py to return an integer and adjusting some tests accordingly
Change: 117834924 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,666,831,000 |
4895e65c156a0581ac29e9605b8704d6b3b53324 | Remove size = "medium" default from tensorflow.bzl functions
We want explicit size arguments for all tests.
Change: 117835174 | Geoffrey Irving | geoffreyi@google.com | 1,458,666,945,000 |
0f533f9fc1d334616ca3a029403f5aa5e88fc77c | Set size for remaining tensorflow tests
Change: 117839211 | Geoffrey Irving | geoffreyi@google.com | 1,458,668,965,000 |
e7668c6a920c77bd77e91f71db0465edcc75e3a6 | Re-arranging tensorflow contrib targets.
Change: 117849583 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,674,021,000 |
7d6bff455d73aca56adcd5212e1b8dbc94e7360a | fixed typo in website
Change: 117854099 | Olivia Nordquist | nolivia@umich.edu | 1,458,676,706,000 |
4bc8285bb65da03862f4131bf7b4df6fe32de560 | Code cleanups as per style conventions.
Change: 117859536 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,679,677,000 |
c51863f3b705e4d285e7c29487aaa70d657bead1 | Add __pycache__ to .gitignore | Geoffrey Irving | geoffreyi@google.com | 1,458,682,916,000 |
da16fd688b23d55422b7ed3a6f196991931278ba | Added a note about compiling user op library with gcc 5 | Manjunath Kudlur | keveman@gmail.com | 1,458,683,242,000 |
ebfebf595a509766fc22bd83cfff70292cdd8d24 | Merge pull request #1584 from keveman/master. | Manjunath Kudlur | keveman@gmail.com | 1,458,683,644,000 |
fab3bd9b2bc3bb9f44ecc4e96b97081d4dd4d8d2 | Test tf.complex_abs in cwise_ops_test.
Change: 117862980 | RJ Ryan | rryan@alum.mit.edu | 1,458,681,333,000 |
5c4a6481b0556148c12939e55a775a10d9f76a79 | Fix range error from python3 on Mac.
Change: 117865583 | Xiaoqiang Zheng | zhengxq@google.com | 1,458,682,629,000 |
26d92f5a597a84afe75a05cf607d03035b5f2f17 | Add optional Android demo termination after running fixed number of inferences passes, to make benchmarks more reproducible.
Change: 117867195 | Andrew Harp | andrew.harp@gmail.com | 1,458,683,476,000 |
12533dd96379c23a9a1ae47a910b97b058d6fa61 | Reduce conv_ops_test flops by 1000x and unshard
conv_ops_test is currently testing exactly the kernel sizes of Inception 2015.
This doesn't add anything in terms of unit test strength, requires sharding,
and generally makes testing slow. Dividing each of the depth sizes by 10
reduces the number of flops by 1000 and s... | Geoffrey Irving | geoffreyi@google.com | 1,458,683,688,000 |
0663ed111abce98a22d8311a122313df11acde68 | Shrink pooling_ops_test and unshard
pooling_ops_test is only slow because it uses unnecessarily large sizes. Speed
it up by dividing all the depth values coming out of Inception by 30.
After this speedup, there is no need to shard.
Change: 117867788 | Geoffrey Irving | geoffreyi@google.com | 1,458,683,791,000 |
b3da7b11f978f105af0952dd0b1f0fdd92484ded | slice_op_test.py needs size = "medium"
Otherwise it is flaky.
Change: 117868217 | Geoffrey Irving | geoffreyi@google.com | 1,458,684,024,000 |
c8fd4f19ba3c0c8a70f86d65b283e7486b75fc5e | Add open-source CI testing of TensorBoard frontend code via Dockerfile.tensorboard
Adds two new scripts:
- install/install_tensorboard_dependencies.sh which installs e.g. nodejs and chromium-browser
- builds/tensorboard.sh which uses npm to install testing dependencies, uses gulp to compile all the TensorBoard code, a... | Dan Mané | danmane@gmail.com | 1,458,685,607,000 |
d7cf62bbdcfe8806c42421925db6938eb365d1de | Remaining sharded kernel_tests fit in size = "small"
Change: 117871897 | Geoffrey Irving | geoffreyi@google.com | 1,458,686,071,000 |
701792e4d4ecc1597b2f66fc3e63b38b3482c1bf | Fix tf.complex_abs gradient test to not be flaky
Change: 117872871 | Geoffrey Irving | geoffreyi@google.com | 1,458,686,661,000 |
eed87c923ead02c0e209d0b8d2e4c29baa983ffe | Rollback of "Re-arranging tensorflow contrib targets."
Change: 117876000 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,688,402,000 |
884763c04240f533dd059a3e47a24ae465dd777a | Update Eigen NN headers in staging repo to match public contents. | Vijay Vasudevan | vrv@google.com | 1,458,688,627,000 |
0a012801934f6b4ea55ad60e24a09d25dcbaca91 | Merge commit for internal changes | Vijay Vasudevan | vrv@google.com | 1,458,691,132,000 |
1830decf78aa0f3a5b2e467fa7436902734b272c | Updating ops history to reflect new ops. | Vijay Vasudevan | vrv@google.com | 1,458,691,374,000 |
900519d820b7f43d9692a092ab3e667b7630787f | Revert "Add GPU support to Variable and Assign for all POD types"
This reverts commit a8955c165929514f5ef6268010cbe198e4a32164.
I believe this breaks some skflow tests due to a bug in the placer. | Vijay Vasudevan | vrv@google.com | 1,458,693,423,000 |
297c00af36ffd2dccf28eaab3b544f43228f213d | Change shard_count for sharded tests back to 4. | Vijay Vasudevan | vrv@google.com | 1,458,694,348,000 |
95e06c5fea9e730155196f42dea94ece2f56eb62 | Merge pull request #1585 from vrv/branch_117876000
Upstream changes from internal for March 22 | Vijay Vasudevan | vrv@google.com | 1,458,694,975,000 |
f4afb66967a43eb3cd25784133a8160f75ba9050 | Adding optimizer_loss op to layers to provide a convenient way to add training operation.
Change: 117881415 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,691,639,000 |
b7babfd20daabe2b50a34851628e629a4c69df51 | Added standalone kernel_library and gen_op_wrapper_py targets
for sdca_ops.
Change: 117884415 | Manjunath Kudlur | keveman@gmail.com | 1,458,693,635,000 |
de15fe5617ed56d87367436f513d03c8c88e0536 | Delete superfluous file added by a commit | Vijay Vasudevan | vrv@google.com | 1,458,706,026,000 |
b3d862b376656c0f58b729e1848ea6e0c5bebf2a | Merge pull request #1589 from tensorflow/vrv-patch-1
Delete superfluous file added by a commit | Vijay Vasudevan | vrv@google.com | 1,458,708,128,000 |
d56b1968633a544a17c63b216262acecf9b5793f | Rollback of "Add GPU support to Variable and Assign for all POD types"
Previously only float and double variables were allowed on the GPU. Now
anything other than string works. The same goes for Assign. AssignAdd
and AssignSub have been left alone for now.
Change: 117887590 | Martin Wicke | wicke@google.com | 1,458,696,023,000 |
4cec6830bc0aaed6cc089c06c437eae1f240758b | Add extension points to the gRPC Session and Server code.
This enables a subclass of GrpcServer or GrpcSession to override the
server credential and channel creation logic, allowing the use of
different security mechanisms.
Change: 117900179 | Derek Murray | mrry@google.com | 1,458,711,455,000 |
18cbeda07a526acbf899ac2363541b8f0b6df29a | Adjust supervisor.py to "bool strictness"
Change: 117900461 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,711,899,000 |
37606a4c63364c56a0834d281023b62d2bda6cd8 | Merge changes from github, some fixes to adhere somewhat
to our requirements for skflow.
Change: 117901053 | Vijay Vasudevan | vrv@google.com | 1,458,712,890,000 |
ed121aa170011aa15569670e6bb7090787dd60f6 | Merge commit for internal changes, including
skflow import fixes. | Vijay Vasudevan | vrv@google.com | 1,458,714,726,000 |
60dcb469fb81430c6a8d0cfe45174e761bf3c31d | fix test sizes | Jan Prach | jendap@google.com | 1,458,717,646,000 |
94c512f30fb2104b448b08707fef43d6c09949ab | Fix a missing-space issue in with_the_same_user.sh | Shanqing Cai | cais@google.com | 1,458,743,189,000 |
d60904080994559876ba93fbc2c2f26aaafb886c | Merge pull request #1591 from jendap/fix-test-duration
fix test sizes | Geoffrey Irving | irving@naml.us | 1,458,745,307,000 |
ed0c46099e578c59f3c3172a1879239ebff7dea3 | Merge branch 'master' into fixes-1 | caisq | cais@google.com | 1,458,746,720,000 |
d144474df0f94d824e2a3dbc56d3d36646853b37 | Merge pull request #1596 from caisq/fixes-1
Fix a missing-space issue in with_the_same_user | caisq | cais@google.com | 1,458,747,722,000 |
ea795f46ed4abd1ce9b3f5fd09622ab84539a27e | Merge branch 'master' into fix-cross-entropy-docs | Robert DiPietro | rdipietro@gmail.com | 1,458,748,701,000 |
c936c06a20b867eaa868e9f2c4fb53856391b18f | Merge pull request #1590 from vrv/branch_117901053
Merge internal changes from late March 22nd | Vijay Vasudevan | vrv@google.com | 1,458,753,225,000 |
04b0babb5c0825e0341615b54d9cf69e83714979 | Merge pull request #1580 from rdipietro/fix-cross-entropy-docs
Fix docs for cross-entropy loss functions | Vijay Vasudevan | vrv@google.com | 1,458,753,297,000 |
127fe199cc65c017d3f644b1ab4d7bdcd535eb32 | finished testing 2LSTM and 2BasicLSTM | Vu Pham | phvu225@gmail.com | 1,458,755,153,000 |
c1707650883996c928a6c6487066f8bf59b45629 | Merge branch 'master' into enhancement/grid-rnn | Vu Pham | phvu225@gmail.com | 1,458,755,354,000 |
b885fe41e52dac6fbfc4fab36a5d2a4b61973687 | Add %matplotlib inline on the first cell
Config the matlotlib backend as plotting inline in IPython | lekaha | lekaha@gmail.com | 1,458,752,837,000 |
35c3f16f81b6a1d39a004c97160c22b73d3798e5 | Merge pull request #1599 from lekaha/add_matplotlib_inline. | Vincent Vanhoucke | vanhoucke@google.com | 1,458,755,560,000 |
60dfe8852cfd9565e622e423098bdad746dc3aee | Make it possible to set a data type for variables that come out of constants
without major contortions, e.g.
foo = variables.Variable(3.14, name="pi", trainable=False, dtype=tf.half)
Change: 117916139 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,730,625,000 |
7c7373505a33204deae163feaa835dc91adf1f73 | Print per-node step stats from Android demo for better benchmarking.
Change: 117928611 | Andrew Harp | andrew.harp@gmail.com | 1,458,744,196,000 |
9cd19435910fc902e982e394a4bf3b35c4c60a01 | Print per-node step stats from Android demo for better benchmarking.
Change: 117931327 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,746,334,000 |
0ff906356e49249f19a1060138d7c8eb22e5d4cc | Raise sharded kernel_tests to size = "medium"
cwise_ops_test is still timing out frequently.
Change: 117935269 | Geoffrey Irving | geoffreyi@google.com | 1,458,749,056,000 |
d79bc31f45b1fdf98bdcac022f0a86a3e7fbf860 | Fix a case where DT_HALF would not be part of the right type lists
for __ANDROID_TYPES_FULL__.
Change: 117936401 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,749,782,000 |
9394661924a37fe17e5f11f43833e9bbffbf23ff | Fix dense_shape check.
This rectifies the following error:
TypeError: Using a `tf.Tensor` as a Python `bool` is not allowed. Use `if t is
not None:` instead of `if t:` to test if a tensor is defined, and use the
logical TensorFlow ops to test the value of a tensor.
when the conditional branch contains a tf.IndexedSl... | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,750,621,000 |
7323103c1c067ba776a6acf25af82ffbec1290ed | Move the NeuralNetwork code out of third_party/eigen3 and into
tensorflow/core/kernel.
Change: 117941211 | Benoit Steiner | benoit.steiner.goog@gmail.com | 1,458,752,851,000 |
7d471c23a4ea5ac163b7ebc5a137f9650d232348 | Fixes warning in tf.Supervisor by passing a Graph to the summary writer.
This enables shape visualization by default in the graph visualizer.
Change: 117941335 | Derek Murray | mrry@google.com | 1,458,752,912,000 |
0417e6716fb230ab7182325fd58c8b2c090c593d | Modified bench_norm_benchmark to use the new Benchmark API.
To run it now:
bazel run //tensorflow/..:batch_norm_benchmark -- --benchmarks=.. --use_gpu={false/true}
Also a tiny file naming bugfix to run_and_gather_logs_lib.
Change: 117941756 | Eugene Brevdo | ebrevdo@gmail.com | 1,458,753,130,000 |
3ad7d3e289aa8e7af4ac98f7f076a8637dc7ecf8 | Adding an FFmpeg install to the Ubuntu Jenkins test machines. This change
will allow ops to be added that use FFmpeg. Note that OSX machines will be
used to validate TensorFlow without FFmpeg installed.
Change: 117942636 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,753,591,000 |
334702e19a920ac21fbbbf5b14f7619cb860c427 | Reformat __future__ imports.
Change: 117943734 | Martin Wicke | wicke@google.com | 1,458,754,160,000 |
cde9ac2140afd8359c9dbbb169996f1ec9fe43c2 | Fixed a bug related to the gradient of nested conds.
Change: 117949288 | Yuan Yu | yuanbyu@google.com | 1,458,757,003,000 |
e37fa0cfa557a46356b654201e827ec84fd40997 | Add support to synchronize replicas in distributed training
Change: 117950091 | Jianmin Chen | goog.jmchen@gmail.com | 1,458,757,426,000 |
f10637b372b3216400129b42abacd4fe50a6d7ea | Allows explicitly setting a gradient function for a user defined function.
Change: 117950844 | A. Unique TensorFlower | nobody@tensorflow.org | 1,458,757,828,000 |
2966fcdae0cf7d75f4eca027f3003e355089b0ed | Added Python APIs for inspecting checkpoint content.
DebugString()
HasTensor()
GetVariableToShapeMap()
Example use can be found in PywrapTensorSliceReaderTest in
third_party/tensorflow/python/training/saver_test.py
For example, To use (in Python):
try:
reader = tf.train.NewCheckpointReader(your_checkpoi... | Sherry Moore | sherrym@google.com | 1,458,758,408,000 |
20a8d1ff2c320dfa46df2bab05334b080aff2444 | Don't silently clip the range argument to int32 before
storing it internally as an int64 (improve future compatibility
with sampling from huge tensors. Note that this CL isn't enough
to permit huge tensors, because we also clip their size at
other points in the process, but it's a step forward.)
Change: 117953355 | David G. Andersen | dga@google.com | 1,458,759,230,000 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.