hash stringlengths 40 40 | msg stringlengths 1 131k | author stringlengths 1 33 | email stringlengths 0 57 | date int64 1,447B 1,698B |
|---|---|---|---|---|
8eaf671025e8cd5358278f91f7e89e2fbbe6a26b | Merge pull request #790 from dongjoon-hyun:print_correct_host_information. Closes #790 | Martin Wicke | martin.wicke@gmail.com | 1,453,010,226,000 |
203723bb7d7718ba6cefdfbd044595ea9af95225 | + Fixed inconsistent sized batch and added note | terrytangyuan | terrytangyuan@gmail.com | 1,453,087,910,000 |
02fdbf45bf7e2c024ed75e37554950804fd5c99b | + one-hot encode per class for dask feeder and re-wrote tests | terrytangyuan | terrytangyuan@gmail.com | 1,453,093,621,000 |
a8132bdcbc014234d4cd103ecffb1ddd9bc64a1a | + Check for pd.Series inferred from sampled out and fit is successful | terrytangyuan | terrytangyuan@gmail.com | 1,453,094,988,000 |
9df70033b71e7a4fa810fdfc71a3715efe9be542 | * DaskDataFeeder works with iris with same accuracy | terrytangyuan | terrytangyuan@gmail.com | 1,453,096,316,000 |
48a95edc43fa420342c8002b3643afbbbffb0065 | Improve gradient support for nested while loops and conditionals.
1. As a result, the gradient code for while loops has been restructured (and simplified).
2. Optimized the handling of "unused" loop variables of a while loop in gradient computation. Gradient computation will only be performed for loop variables that contribute to ys for gradients(ys, xs).
Change: 112333290 | A. Unique TensorFlower | nobody@tensorflow.org | 1,452,983,797,000 |
8ffdf4a33941e8b1f259c3834d4376cd1cdf3790 | Moved the flag parsing into a separate module.
Tidying up the label_image example so that the argument handling is taken care
of outside of the main flow of the code, to make understanding it easier.
Change: 112333770 | Pete Warden | pete@petewarden.com | 1,452,985,143,000 |
fa7e6dd73a0dd5548d263322341ba7a74092f173 | Make tensorflow/core:lib public.
Change: 112342374 | Josh Levenberg | josh11b@tensorflow.org | 1,453,008,769,000 |
2e65c5c9a4a828e3fd1ee51e0965e5add433766d | Tool script for converting checkpoints into 'frozen' GraphDefs where variables are replaced by Const ops.
Change: 112363668 | Pete Warden | pete@petewarden.com | 1,453,064,096,000 |
6f62e435ab6c36dfdfdef1acd580b5f278f6723c | Adds enough auditing to make it possible to track tensor buffers throughout an execution, and build a cost model of memory usage.
There are two main components:
1) GPU allocators now assign to each allocated tensor buffer a unique ID so its use can be tracked within and across steps.
2) The checkin cleans up the tracking of usage of Tensor buffers, and makes it work for both sync and async kernels (async kernels did not previously track gpu memory correctly). Each use is now tracked by the OpKernelContext (for allocators that need this support) in a single uniquified set of TensorReferences. When the kernel finishes, the executor retrieves the list of references, logs it if needed in the nodeexecstats, then passes it to the device, which may add an additional reference to keep the memory from being reused until the execution completes. When the tensor is logged in the nodeexecstats a flag is set if there is a single remaining reference to the buffer, which means that the memory will be freed once the Op completes.
Change: 112375683 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,088,429,000 |
a84a81a7379507f8fcdd0d6118afc2d5044d159e | * Supports nested function calls;
* Supports uses python class method as a tf function prototype.
Derives the generate function name from the class name and the method name.
* Changes one LOG(INFO) to VLOG(1), which is too verbose.
Change: 112384148 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,102,306,000 |
80a481b508d4e63066d45e9fe1c9bee0090299d3 | Unspecified message. | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,141,810,000 |
21277cec3a85ee4572ffbc8e21937eb0c35d07be | Unspecified message. | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,149,076,000 |
1f7947ab9dbf725b38385be6405ae6040d3b522b | reduce logging level of a useless message
Change: 112422453 | Yaroslav Bulatov | yaroslavvb@google.com | 1,453,149,357,000 |
ad05c77b8bde8ed47da86dad698bf77c301cf8fb | Added the missing build_config:gtest dependency to command_line_flags_test
target.
Change: 112423695 | Manjunath Kudlur | keveman@gmail.com | 1,453,151,405,000 |
04bad79a588a09f7642544c5e1cc168f7dfbcab4 | Fix examples/label_image/...
- string -> tensorflow::string
- Depend on core:test_main for the main function for test.
Change: 112424835 | Manjunath Kudlur | keveman@gmail.com | 1,453,153,539,000 |
8a445e54b893f313ca44c7ef124b1884f4994e95 | DEFINE_bool -> DEFINE_boolean
Change: 112425239 | Manjunath Kudlur | keveman@gmail.com | 1,453,154,405,000 |
00440e99ffb1ed1cfe4b4ea650e0c560838a6edc | Merge commit for internal changes | Manjunath Kudlur | keveman@gmail.com | 1,453,154,481,000 |
d04e3b7fdf9e6bd7a6c0a7b20b69d52d890b7f51 | Fix typos in core/framework. | Dongjoon Hyun | dongjoon@apache.org | 1,453,112,104,000 |
b63c57620d08385975a22f2784e1006af9e17d78 | Merge pull request #801 from dongjoon-hyun/fix_typos_in_framework. Closes #801 | Vijay Vasudevan | vrv@google.com | 1,453,157,191,000 |
77a3d2c2fa881511e69d33e02451c49b5d53c8b7 | * prediction on each partition | terrytangyuan | terrytangyuan@gmail.com | 1,453,169,867,000 |
aeefb1ec2417a79ebc05943a7e3bcd3f25eb116d | + Example of out-of-core dataframe classification | terrytangyuan | terrytangyuan@gmail.com | 1,453,171,255,000 |
c082ae91ef20c0ef7566b2555960ba3b4bb1e9f5 | * Modification of out-of-core example | terrytangyuan | terrytangyuan@gmail.com | 1,453,171,642,000 |
8a1c308450b64a7d3c2ca520109c4909acbf13ff | Remove word2vec's dependency on RE2. This is primarily to allow inputs
over 2GiB in size.
As part of this, add ConsumeNonWhitespace() to str_util.
Change: 112436569 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,174,670,000 |
0de9d7b4757e36fb1e1c5a22c16965afa15c086e | Merge commit for internal changes | Manjunath Kudlur | keveman@gmail.com | 1,453,217,717,000 |
7ef2351fc34a67e4c4af70a41143199712da7533 | ci_build: siplify interactive mode | Jan Prach | jendap@google.com | 1,453,091,034,000 |
4322ce2450d43268f1263bfdbc32013b54821826 | ci_build: remove common section from .bazelrc
* otherwise bazel fail on parameter unsupported by the command
(running "blaze query" would fail before) | Jan Prach | jendap@google.com | 1,453,091,820,000 |
e7183d14233ffd07c33b08d893a7edf548ac89c6 | ci_build: add support for python3 build | Jan Prach | jendap@google.com | 1,453,242,997,000 |
ec00878777f253c79b5eda872be4116fb5d64fef | Merge pull request #815 from jendap/ci_build. Closes #815 | Martin Wicke | wicke@google.com | 1,453,244,153,000 |
e0a6a3d620a585ca5e9bfd50f98f80b1eada52e4 | Further improved the performance of the contrast adjustment code by
enabling nvcc to optimize the broadcast of scalars | Benoit Steiner | benoit.steiner.goog@gmail.com | 1,452,907,707,000 |
04de87f1204103cf9b753858ecae33099aafb4c2 | Merge pull request #814 from benoitsteiner/master. Closes #814 | Vijay Vasudevan | vrv@google.com | 1,453,249,616,000 |
e10333872630028e4c0a90d1facedd0ef8081b35 | * Minor modification in out-of-core example | Yuan (Terry) Tang | terrytangyuan@users.noreply.github.com | 1,453,261,469,000 |
fd7fe21131b2100eecb90eb3afa76c43720dae94 | Fixed the usage of metrics.accuracy_score(). | okoriko | eurico@phybbit.com | 1,453,267,546,000 |
1925f1976c7d9d102fe76844815a309da1f7ea04 | ignore install six for tensorflow to get 2.7 work | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,268,236,000 |
6bb6736e937729f18ffa8be75e1f50a90aa3003e | Tests passed. | makseq-ubnt | makseq@gmail.com | 1,453,303,350,000 |
618efa26ac0aed2b07c35c13bbe0bc36abf93b17 | Unittests changed. | makseq-ubnt | makseq@gmail.com | 1,453,303,546,000 |
f592f23775e2a6ac75496829db5005d3bb70a3d2 | Replacing reference 'names' variable with 'example_names' variable.
Change: 112481326 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,224,720,000 |
e39629219e748b08177f2c457ba45d51f5370aae | PaddingFIFOQueue is like FIFOQueue but allows dynamic shapes (using padding with DequeueMany)
Change: 112482056 | Eugene Brevdo | ebrevdo@gmail.com | 1,453,225,174,000 |
91facc371f5ec5abb91b9b921fb44428ea8ff939 | Add missing license header from unique_tensor_references.h
Change: 112490207 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,229,546,000 |
757eaa78fe3de8193f330829a6d0ec8f7220d5e0 | Fix comment typo.
Change: 112490296 | Josh Levenberg | josh11b@tensorflow.org | 1,453,229,583,000 |
ee274fdd0af515a09a3a45a1d31065b3edd06087 | Clean up GPU kernel registrations for control flow ops.
Change: 112492676 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,230,708,000 |
2712ed6de036e16b2599fcab2071acd7bbf8b17a | Implement TensorArray forward ops.
Allows dynamic writing to- and reading from- an array of Tensors (size of the array determined at run time).
This is useful for, e.g., While loops. Each while iteration can write to the Array; and the final handle can be used with Concat to get all the outputs in one Tensor.
No gradient support yet, this will be implemented in a future CL.
Change: 112493043 | Eugene Brevdo | ebrevdo@gmail.com | 1,453,230,910,000 |
ed5205ae9bcd11b0d7b9dcb9aea74d500ac61d8b | Simplify ConcatGrad so that it generates N+2 nodes instead 4N+16 nodes.
Change: 112499672 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,234,277,000 |
b916d6591318544d04fec690dcd944756cc1c300 | Change: 112499684 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,234,282,000 |
7ffa2d77111d75589f37ecf0ec6e8ee1fe53b995 | Add support for Refs in control flow ops (Merge, Switch, Enter, Exit).
Change: 112505342 | Eugene Brevdo | ebrevdo@gmail.com | 1,453,237,289,000 |
c8eaac926c929e07ac8db69f67803a2223ff2d93 | Many tensorflow/core build clean ups.
Change: 112523833 | Josh Levenberg | josh11b@tensorflow.org | 1,453,247,567,000 |
13958b24e3c982f826ad5e2e4d044c496f947d6b | pylint check fixed. | makseq-ubnt | makseq@gmail.com | 1,453,306,098,000 |
9d5d3d071bb50d09060d020cca633202333c2a58 | ci_build: fix bash array usage | Jan Prach | jendap@google.com | 1,453,311,113,000 |
5f7ecbc35d085861acdff75d030e34656202cb0e | Merge pull request #820 from jendap/ci_build. Closes #820 | Vijay Vasudevan | vrv@google.com | 1,453,311,601,000 |
a7b25ede1933c22d43057827eefafc0abb610fbb | Fix syntax error when executing setup.py using python 3.4
There are SyntaxError messages during processing reader.py and
cwise_ops_test.py due to incompatible syntax of python 2 and 3.
According to 'PEP 3113 -- Removal of Tuple Parameter Unpacking',
a tuple parameter is deprecated from 3.0 version of python, so we
modify method _build_vocab to make it compatible with python 2 and
python 3. | Bi-Ruei, Chiu | biruei.chiu@gmail.com | 1,452,960,649,000 |
1abdddb84798f4813e5e6037bae53090de7b9fe6 | Merge pull request #792 from brchiu/master. Closes #792 | Vijay Vasudevan | vrv@google.com | 1,453,311,668,000 |
933c08796a0d3fd570c5c0b7aa5aa81d3c57370a | Use the most specific possible shape in the merge op.
Add tests for the new merge op shape inference.
Speed up the bidirectional RNN by providing the sequence length to each RNN.
Add tests for the bidirectional RNN when the sequence length or shape is not provided. | Kenton Lee | kentonl@cs.washington.edu | 1,452,909,069,000 |
96289f712558a64a957a80b0465341fef10d90ab | Merge pull request #738 from kentonl/birnn_fix. Closes #738 | Vijay Vasudevan | vrv@google.com | 1,453,311,996,000 |
8dbf9b0923642e902450c47d1a970452380f07b4 | fix docs of 'Saver.max_to_keep' | Linchao Zhu | zhulinchao7@gmail.com | 1,453,290,254,000 |
90fb80a7b3ec90339ed82234a99de2add0952d98 | Merge pull request #818 from ffmpbgrnn/master. Closes #818 | Vijay Vasudevan | vrv@google.com | 1,453,312,057,000 |
c878ce6ae549a72dbdbd2c41f33fa3dec75bb555 | Change DebugString behavior to ShortDebugString
The old behavior of DebugString is needlessly verbose and is quite confusing
for scalar shapes (it produced the empty string). Now DebugString is the same
as ShortDebugString.
A future commit will remove ShortDebugString.
Change: 112590646 | Geoffrey Irving | geoffreyi@google.com | 1,453,312,401,000 |
f6f1bb541329ab2b3315d864252e6daece5e3187 | update docs
Change: 112591828 | Martin Wicke | wicke@google.com | 1,453,313,043,000 |
585983b56648f066c05f3cb78fdadc339efc3b15 | Change: 112595580 | Manjunath Kudlur | keveman@gmail.com | 1,453,315,022,000 |
7a73630e276e5bc36d89d75cfdd33d1d3a5f865b | Add missing arguments.
* For `docs.py`, add `name` for exception message generation.
* For `rnn_cell.py` and `rnn_test.py`, add `scope=None` for inheritance. | Dongjoon Hyun | dongjoon@apache.org | 1,453,187,133,000 |
c001b0de97f0c30132ada095f6bf4ce127c2e1ba | Merge pull request #807 from dongjoon-hyun/add_missing_argument. Closes #807 | Vijay Vasudevan | vrv@google.com | 1,453,317,936,000 |
cf171aa47573d3ac6298b59769d0e9c41d76fcc1 | Fix typos in comments. | Dongjoon Hyun | dongjoon@apache.org | 1,452,979,460,000 |
4c3b060b9b063a247172c939e0e1314577b581e3 | Merge pull request #825 from dongjoon-hyun/fix_typos. Closes #825 | Vijay Vasudevan | vrv@google.com | 1,453,327,056,000 |
07f75e514f152fa99f610cd81e3b1ba8c13f63ad | Remove mutable default value.
Up to now, it's not a problem since `unigrams` is always given.
However, it might cause a problem in the future because
a mutable default parameter value is created once and shared.
According to the comment, use '()' instead of 'None'. | Dongjoon Hyun | dongjoon@apache.org | 1,453,192,420,000 |
8c95ea4cf1324d3911452609f93f3de0e64cd9d1 | Merge pull request #809 from dongjoon-hyun/remove_mutable_default_value. Closes #809 | Vijay Vasudevan | vrv@google.com | 1,453,331,043,000 |
4ebc5401287ea1f31c4c31171c2eb6a3437a8442 | Add --spawn_strategy=standalone to os_setup.md
It seems to solve more problems than it causes. Also fixes some formatting/grammar nits. | Martin Wicke | martin.wicke@gmail.com | 1,453,333,845,000 |
8edd318ce6e45cc03811d999c4958e63e8d6d39b | Merge pull request #826 from tensorflow/os-setup-spawn-strategy. | Martin Wicke | martin.wicke@gmail.com | 1,453,336,813,000 |
47dc8be96e3b6d7fef9c15e2bb1fe9e7444ef8f5 | Rearrange BUILD into sections, and in particular put the targets in
the public section in a sensible order (smallest to largest, matching
the documentation comment).
Plus:
* Update documentation to reflect that test_main is not public.
* Remove "friends" package_group now that it is unused.
Change: 112605117 | Josh Levenberg | josh11b@tensorflow.org | 1,453,320,167,000 |
084075e4e967677fc0be9fe921a0157126e69617 | More function support.
Change: 112611228 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,323,914,000 |
1846063e694f28b2da3928c7db6ad8ae7e4452f3 | Removes some unintended change.
Change: 112611994 | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,324,398,000 |
3ea040538b318b8e2d76f626fa34df801e1af1d0 | Adds `tf.decode_json_example()` op that parses JSON into binary Example protos.
Change: 112615357 | Derek Murray | mrry@google.com | 1,453,326,462,000 |
6cc49690efd9b1017dc9b7b3659ec02431fca490 | Improve exception message when `TensorShape.with_rank()` fails
Previously this would pass on the exception message from
`TensorShape.merge_with()`, which is cryptic for users who don't (and
shouldn't need to) understand how shape inference works. This would
arise, for example, in the error message for `tf.matmul()` when passed
a non-matrix, as noted here:
http://stackoverflow.com/questions/34908033/tensorflow-exception-with-matmul
Change: 112621185 | Derek Murray | mrry@google.com | 1,453,329,627,000 |
db7478e8998f7703c57a75a950c905ec0cb59d7b | First step of moving files out of tensorflow/core/public/. Here
we copy the original files to their new location and make the public/
versions #include the new location. Once all references are updated
to point to the new location, we can delete the originals in public/.
Change: 112622561 | Josh Levenberg | josh11b@tensorflow.org | 1,453,330,490,000 |
877fcd1a113797a1c5847dd5fdbef7868addded0 | Prepare to hide tf.tensor_util
1. There is a new tf.unsupported module to hold things which some people use
but which we don't yet support.
2. tf.tensor_util.ConstantValue is now tf.unsupported.constant_value. Most
users use this, but tf.tensor_util.ConstantValue is still available; it
will be removed in a following commit.
3. tensor_util.MakeTensorShapeProto is now make_tensor_shape_proto. It looks
like all users of this access the tensor_util module directly (not through
tf), so for now it is not in unsupported.
This commit does not remove tensor_util from tf.__all__; a few more downstream
users must be changed before that can happen.
Change: 112626961 | Geoffrey Irving | geoffreyi@google.com | 1,453,332,966,000 |
e62d685b9e9b0d54d0f9ab36aa582715f5e076af | Change: 112636955 | Manjunath Kudlur | keveman@gmail.com | 1,453,339,081,000 |
2d7b6485799802d4381cc8353617f3d2651f5530 | Unspecified message. | A. Unique TensorFlower | nobody@tensorflow.org | 1,453,340,878,000 |
ba1975f067062a38dc5e760f7782408cf52c87d7 | Merge pull request #86 from Phybbit/master
Consistent usage of metrics.accuracy_score with the documentation | Yuan (Terry) Tang | terrytangyuan@users.noreply.github.com | 1,453,341,247,000 |
6bf5a24e8805c57a2d7e5741519a090228c76a89 | Change: 112640197 | Manjunath Kudlur | keveman@gmail.com | 1,453,341,816,000 |
20882f501cb218b359ba9f20786a2217e68cc436 | Change: 112641949 | Manjunath Kudlur | keveman@gmail.com | 1,453,343,205,000 |
e4029435aefcc6c406c9a485bcb7a0893a95f28e | Rollback of "Updated TensorBoard logic for discarding events after crashes to discard out of order events only within a particular tag. This was changed because race conditions in supervisor was causing many events to be unintentionally discarded."
Change: 112644077 | Vijay Vasudevan | vrv@google.com | 1,453,345,680,000 |
cf6cdcf6dc7e5dba2b0f729b86bf1f14ca73666e | Change: 112644520 | Manjunath Kudlur | keveman@gmail.com | 1,453,346,222,000 |
307c35936d3a6a931dca98bdb7c0ec9c0f904071 | Merge commit for internal changes | Manjunath Kudlur | keveman@gmail.com | 1,453,346,344,000 |
207578cd018e98d0e1cca0df9f146f4cec4b0631 | Fixes #84: Reshaping the labels for regression to compute correctly cost for batch_size>1. Also started working on class_weights. | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,348,651,000 |
18a3e2aa2e5f8caa1b92bd54fe041e10275611f6 | removing debugging code | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,348,862,000 |
6964f2d33b32ee338d2e9e2d0076bcb0311e4c94 | Bumped up the default number of steps to get more training if user doesn't provide steps variable. | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,349,098,000 |
342bdf79151450fde6ab08a20b4f5f00c18e8c25 | Updated tests to use tf.test.TestCase instead of googletest. Added test_regression for recreating weights | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,350,120,000 |
d4b81b2339e259dcd9f90927d4c4799df7d34d19 | + rnn_estimator | terrytangyuan | terrytangyuan@gmail.com | 1,453,346,845,000 |
6303920180af729d4421c37300c8f0ce0ccf1c9a | + RNNClassifier and RNNRegressor | terrytangyuan | terrytangyuan@gmail.com | 1,453,347,514,000 |
f86793691de747a430244beb5c6423e7873285f8 | + input_op_fn for rnn_estimator | terrytangyuan | terrytangyuan@gmail.com | 1,453,347,618,000 |
ed2f2f27ba23d71ac97ddf5e6fb14dcde2a3d39f | * minor changes | terrytangyuan | terrytangyuan@gmail.com | 1,453,348,122,000 |
81cef20c021492eabbaa483e5c7ac6d9224bd029 | - revert last commit | terrytangyuan | terrytangyuan@gmail.com | 1,453,348,189,000 |
e35b6a98d91a29e0b21dd077347d4ca22067833b | * Fixed all lint/nosetests errors | terrytangyuan | terrytangyuan@gmail.com | 1,453,350,800,000 |
a15b57dfa31556a0de0ed01d1c7265575de31a85 | Added weights and biases properties to DNN estimators | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,351,839,000 |
db2ce0d4646461e4abfac399052f4723f9f17a0b | Updated readme - added more about types of models and bumped min requirement for TF to 0.6 | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,352,019,000 |
5966fc51b8f9bc8a8a3682005bb9151d7f5f0dde | More bolding and formatting in README | Illia Polosukhin | ilblackdragon@gmail.com | 1,453,352,070,000 |
e829124f12064691271f8f5f321222fe6d23ef22 | + Example of using built-in rnn model | terrytangyuan | terrytangyuan@gmail.com | 1,453,352,463,000 |
783cef9e7bba16307c3a246df37b9460f27ee5d4 | + Tests for rnn models | terrytangyuan | terrytangyuan@gmail.com | 1,453,354,642,000 |
86505e92b641c87b2d15bcd28eacdba9430e7e10 | * Fixed failed test due to change of default for steps in linreg since it overfits | terrytangyuan | terrytangyuan@gmail.com | 1,453,354,719,000 |
34a44633d3b2f884edb0c84c500165bdfbb5f9ff | + More tmp coverage for RNN model | terrytangyuan | terrytangyuan@gmail.com | 1,453,355,357,000 |
024c19381232fbae7a010d8b86658fc530712edf | Merge branch 'rnn' | terrytangyuan | terrytangyuan@gmail.com | 1,453,355,420,000 |
9519c7d70f4f0e2dec59467f359c35e7ca9b9e33 | + Default values for rnn models | terrytangyuan | terrytangyuan@gmail.com | 1,453,356,174,000 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.