aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* Allow gradient computation for scalar values in e.g. ReLUGravatar A. Unique TensorFlower2016-07-12
| | | | | | | Fixes #3277. Changes binary ops used to compute gradients for several operations to support zero-rank tensors. Change: 127255053
* Fix typo involving """" where """ would do.Gravatar A. Unique TensorFlower2016-07-12
| | | | Change: 127254774
* Update estimators documentation to mention weighted sparse column.Gravatar A. Unique TensorFlower2016-07-12
| | | | Change: 127254286
* Automated rollback of change 127233960Gravatar Vijay Vasudevan2016-07-12
| | | | Change: 127253427
* Add support for exponential smoothing for TensorBoard.Gravatar Renato Utsch2016-07-12
| | | | Change: 127245277
* Switched to the latest version of Eigen that provides significant performanceGravatar Benoit Steiner2016-07-12
| | | | | improvements for fp16 Change: 127233960
* Automated rollback of change 127146715Gravatar Vijay Vasudevan2016-07-12
| | | | Change: 127225742
* Fix tf.multinomial to not crash on empty inputGravatar Geoffrey Irving2016-07-12
| | | | | | | | | | On the GPU, tf.multinomial uses Eigen. On empty input, this triggers a bug in Eigen causing a crash. Fix this by not executing the kernel in the empty output case. Also fix shape validation assertions to handle more corner cases (hopefully all of them). Change: 127223716
* Avoid extra zero padding of strings in py_funcGravatar Geoffrey Irving2016-07-12
| | | | Change: 127223683
* Adds an interface for asking malloc to release free memory back to the OS.Gravatar Christopher Olston2016-07-12
| | | | Change: 127216353
* Bug fix: initialize WorkerCacheLogger::want_logging_count_ to 0.Gravatar A. Unique TensorFlower2016-07-12
| | | | | | | | | | | Without this, collection of detailed execution timing information would default to being on all the time quite often, resulting in significant performance and memory allocation overhead. Reduces number of allocations for one benchmark from 120M to 101M. Co-discovered by sanjay@ and myself. Change: 127212900
* Disable flaky testsGravatar Vijay Vasudevan2016-07-12
| | | | Change: 127212113
* Made tf.py_func an aynchrounous operation.Gravatar A. Unique TensorFlower2016-07-12
| | | | | | | | | | | | | | | | | | | This enables multi-threading of the tf.py_func operations in cases where wrapped Python functions releases the GIL. The following example now takes 3s instead of 6s: def foo(): time.sleep(3) return 1 def create_py_func(): return tf.py_func(foo, [], [tf.int64]) a = create_py_func() b = create_py_func() session.run([a, b]) Change: 127210255
* Add C++ shape inference for TopK and TopKV2.Gravatar A. Unique TensorFlower2016-07-12
| | | | | | | | | Change InferenceContext: - allow Dim(s, negative_idx) - add helper to get dimension value from scalar input tensor. - allow SubShape(s, start, end, &out). Support negative indexes, and indexes > rank, in SubShape (to match pythonic indexing). Change: 127206793
* Readd the destroy() method, as it was still being used by the code.Gravatar Renato Utsch2016-07-12
| | | | Change: 127202712
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-07-12
| | | | Change: 127198546
* Add a ParameterizedTruncatedNormalOp to eventually replace the current ↵Gravatar A. Unique TensorFlower2016-07-12
| | | | | | | | TruncatedNormalOp. It takes a matrix of batched parameters (mean, stdev, minval, maxval). Once the GPU functor is added, we can eventually use this op to implement tf.truncated_normal with an optional minval and maxval, and support for batches, and remove the existing TruncatedNormalOp. Change: 127196322
* Change DequantizeOp to use eigen for MIN_FIRST case.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127155464
* Move over recent changes to tensorflow_serving/session_bundle to ↵Gravatar Kiril Gorovoy2016-07-11
| | | | | | | | tensorflow/contrib/session_bundle, including: - Type and error checking when unpacking Any messages - Increased session_bundle test coverage Change: 127152552
* Fix a bug in _DNNLinearCombinedBaseEstimator.linear_weights_.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127146715
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127144423
* Switch to the official boringssl git repository with Bazel support.Gravatar Alexey Surkov2016-07-11
| | | | Change: 127144397
* Add a rot90 op to tf.image to rotate images by (multiples of) 90 degrees.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127141730
* Automated rollback of change 127123966Gravatar Vijay Vasudevan2016-07-11
| | | | Change: 127141513
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127139520
* Fix a bug when loading checkpoints that use relative path.Gravatar A. Unique TensorFlower2016-07-11
| | | | | | Also add additional checks that can be triggered when users manually edit the checkpoint file and can help them debug potential mistakes they may have made. Change: 127138923
* Merge the BaseChart and LineChart classes into one, as the BaseChart is notGravatar Renato Utsch2016-07-11
| | | | | used for multiple charts anymore. Change: 127135452
* Add Bernoulli-Bernoulli KL, remove extra op in log_pmf broadcasting.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127132058
* Explicitly use integer-location based indexing for Pandas objects, which ↵Gravatar A. Unique TensorFlower2016-07-11
| | | | | | | | | | | otherwise default to label-location based indexing. This is necessary because the DataFeeder uses the shape of the objects to generate indices, rather than the actual index. Mixing the two is often benign, but selections, like boolean masks or head/tail slices will cause failures. See: http://pandas.pydata.org/pandas-docs/version/0.17.0/indexing.html#indexing-integer http://pandas.pydata.org/pandas-docs/version/0.17.0/indexing.html#indexing-label Change: 127130473
* Change tensorflow/contrib/session_bundle namespace from tensorflow::contrib ↵Gravatar Kiril Gorovoy2016-07-11
| | | | | | | to tensorflow::serving. Update all Python users of tensorflow_serving/session_bundle/manifest.proto to use the equivalent tensorflow/contrib/session_bundle/manifest.proto since there are otherwise namespace collisions. Change: 127129383
* Add support for Tensors of zero size to unpack.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127127852
* Split out the obsolete histogram chart from tf-chart, so we can refactor the ↵Gravatar Dan Mané2016-07-11
| | | | | | line chart on its own. Change: 127127152
* Increase the error threshold for a test.Gravatar Xiaoqiang Zheng2016-07-11
| | | | Change: 127126726
* Add docs to monitor.py, along with some minor cleanups for consistency and a ↵Gravatar A. Unique TensorFlower2016-07-11
| | | | | | coupld TODOs. Change: 127126685
* Mention upcoming C API work in the roadmap.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127126500
* Local Response Normalization GPU support via Stream Executor.Gravatar RJ Ryan2016-07-11
| | | | | Includes port of internal Stream Executor support for cuDNN normalization. Change: 127123966
* Adding an environment variable to control whether to use Cudnn with AvgPool:Gravatar Xiaoqiang Zheng2016-07-11
| | | | | TF_AVGPOOL_USE_CUDNN. The default is false for now. Change: 127123594
* Add a pretty printing function to contrib.Gravatar A. Unique TensorFlower2016-07-11
| | | | | Similar to tf.Print but with the ability to print SparseTensors and TensorArrays. Defaults to printing the passed through tensor and to printing out the shape, dtype and name of all Tensors. Change: 127122527
* Added a session wrapper for running with monitors.Gravatar Mustafa Ispir2016-07-11
| | | | Change: 127118204
* When an invalid op is registered, ensure that tf.load_op_library returnsGravatar Suharsh Sivakumar2016-07-11
| | | | | | | | | | | | | | | | a nice error rather than silently generating invalid python code. Fixes: - On linux, a fatal log that should be triggered in C++ seems to be ignored causing the tf.load_op_library call to fail and exit. We fix this by propagating the status from op registration to raise an exception from python rather than print a fatal log. The reason why the fatal log is being ignored on linux is unclear, my theory is some sort of SWIG issue, but online research provided no results. - Op registrations that fail are still added to the OpDef registration. This is only an issue when the Fatal log is ignored, but we fix this anyways. Change: 127114085
* Namei wheel file properly to avoid name changes and pip install issuesGravatar Shanqing Cai2016-07-11
| | | | | | | | | | | | On Linux, we have seen the whl files built from the PIP tensorflow builds change their names for unclear reasons, leading to the need to update README.md and install issues like these: https://github.com/tensorflow/tensorflow/issues/1097 http://stackoverflow.com/questions/33622613/tensorflow-installation-error-not-a-supported-wheel-on-this-platform as well as build issues like: http://ci.tensorflow.org/view/Nightly/job/nightly-distributed-build-server/132/console This CL aims to freeze the whl file naming pattern on Linux to prevent such issues from happening in the future. Change: 127111314
* Update copyright, and accessor for summaries.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127109341
* Replace "Scikit Flow Authors" with "TensorFlow Authors", and remove "pylint: ↵Gravatar A. Unique TensorFlower2016-07-11
| | | | | | disable=g-bad-file-header". Change: 127106850
* Replace stratified_sample implementation to a strictly more efficient ↵Gravatar A. Unique TensorFlower2016-07-11
| | | | | | implementaiton. The new implementation discards fewer data, and scales better based on the number of classes. However, it requires knowing the class distribution of the data. Change: 127104811
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127103847
* Merge changes from github.Gravatar Andrew Harp2016-07-11
| | | | Change: 127101926
* Fix label_smoothing for sigmoid_cross_entropy_loss, and test for ↵Gravatar A. Unique TensorFlower2016-07-11
| | | | | | | softmax_cross_entropy_loss with label_smoothing. The sigmoid cross entropy with label smoothing was broken, and worse the tests for both that and the softmax cross entropy with label smoothing were broken. I've fixed both issues here and added comments walking through the two examples in the tests so as to not inadvertently check in broken tests again. Change: 127100213
* Support weighted sparse column in sdca optimizer.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127093435
* Add a private bit to the upload to graph observatory script and the dataset ↵Gravatar Nikhil Thorat2016-07-11
| | | | | | | | JSON format. If a dataset is private, it will only show up if it is requested via URL, otherwise it will not appear in the dropdown. Change: 127090337
* Internal only change.Gravatar A. Unique TensorFlower2016-07-11
| | | | Change: 127084359
generated by cgit v1.2.3 (git 2.39.1) at 2024-07-06 17:47:06 +0000