aboutsummaryrefslogtreecommitdiffhomepage
Commit message (Collapse)AuthorAge
* C API: Comment tweaks and some suggested corner case handlingGravatar Asim Shankar2016-11-16
| | | | Change: 139349604
* Fix LoadSavedModelFromLegacySessionBundlePath() to return error if ↵Gravatar Li Lao2016-11-16
| | | | | | LoadSessionBundleFromPathUsingRunOptions() fails. Change: 139348875
* Fixing a bug in the MutableDenseHashTable implementation where the ↵Gravatar Rohan Jain2016-11-16
| | | | | | | difference in shapes between the Insert and Import functions was causing issues with a vector key and scalar value input. Fixed by splitting the LookupInterface CheckKeysAndValueTensors method into one for Insert and the other for Import. Change: 139346138
* Remove explicit calls to call_cpp_shape_fn from third_party/tensorflow/python.Gravatar A. Unique TensorFlower2016-11-16
| | | | Change: 139345066
* Do not calculate loss in head if mode==INFER.Gravatar A. Unique TensorFlower2016-11-16
| | | | Change: 139339828
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-11-16
| | | | Change: 139328977
* Adds max_norm option to embedding_lookup (and upstream functions).Gravatar A. Unique TensorFlower2016-11-16
| | | | Change: 139325873
* Embedding Projector: Guard from null strings when adding word breaks.Gravatar Dan Smilkov2016-11-16
| | | | Change: 139324859
* Simplify data parsing in TensorForest. Remove unused ↵Gravatar A. Unique TensorFlower2016-11-16
| | | | | | sparse_values_to_indices op. Change: 139320375
* Migrate TensorFlow to libjpeg-turboGravatar Justine Tunney2016-11-15
| | | | | | | | | This makes JPEG go 2x faster on x86_64 (k8), arm7, and arm8. On all other CPU targets, e.g. x86, JPEG performance should be the same as it was before. Fixes #4807 Change: 139295768
* Set a max number of collected NodeExecStats in StepStatsCollector.Gravatar Suharsh Sivakumar2016-11-15
| | | | | | This fixes an issue where nodes in a while loop were accumulating a very large amount of stats, exceeding the protocol buffer limit. Change: 139277518
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139271215
* Makes the Variable.ref() method private.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139267342
* C API: Helper functions to encode/decode strings for TF_STRING tensors.Gravatar Asim Shankar2016-11-15
| | | | | | | | | Intended to make the use of TF_STRING valued tensors in other language bindings easier. No added unittests since the existing TestEncodeDecodeStrings test case covers the changes. Change: 139263312
* Fix dnn_test, which was cause by a bad sync/merge.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139259251
* enabling tf.round properly!Gravatar Olivia Nordquist2016-11-15
| | | | Change: 139254648
* Add feature to supply SavedModel with a main_op.Gravatar Sukriti Ramesh2016-11-15
| | | | Change: 139254613
* Increase the block size of SwapDimension1And2InTensor3UsingTiles from 32 to ↵Gravatar A. Unique TensorFlower2016-11-15
| | | | | | | 128 threads. This increases the parallelism in this kernel, yielding a nice speedup. Change: 139250945
* BUGFIX: Softplus bijector inverse did not work with small values 'y'. ThisGravatar Ian Langmore2016-11-15
| | | | | change does a 1st order Taylor expansion to solve for small values. Also add a new assert, assert_bijective_and_finite, and corresponding tests for some Bijectors. Change: 139250783
* Embedding projector: Color points based on distance.Gravatar Dan Smilkov2016-11-15
| | | | Change: 139247755
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139247118
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139246371
* Deprecate tf.inv in favor of tf.reciprocal.Gravatar Martin Wicke2016-11-15
| | | | Change: 139240711
* Allow user to close error messages and add indefinite progress bar.Gravatar Dan Smilkov2016-11-15
| | | | Change: 139239348
* Change default for `assign_moving_average` to perform 0-debiasing. This is a ↵Gravatar A. Unique TensorFlower2016-11-15
| | | | | | semantic breaking change. Change: 139237086
* Fixing a bug related to using a string piece after the inner data is destroyed.Gravatar A. Unique TensorFlower2016-11-15
| | | | | The string piece holds a reference to the internal buffer_ when this class gets destructed the buffer gets destructed but the stringpiece can still be used outside (it is an input parameter). The fix is to first copy into scratch (which is an input parameter) and then put the scratch in the stringpiece (result). Change: 139236445
* Removes the _ref dtypes from the public list.Gravatar A. Unique TensorFlower2016-11-15
| | | | | (breaking change) Change: 139235557
* Create constants for all the different kinds of label fill/stroke coloringGravatar Charles Nicholson2016-11-15
| | | | | | | (selected, neighbor, hover). Still black + white pending design. Refactor adapter to use new method styleStringFromPackedRgba, which does the array indexing and string construction. Change: 139235290
* Wrap checkpoint and metadata file paths.Gravatar Dan Smilkov2016-11-15
| | | | Change: 139229687
* In call_cpp_shape_fn, use whether the shape inference fn requested input tensorGravatar A. Unique TensorFlower2016-11-15
| | | | | | | | | values to rerun the shape inference function. This removes the need to pass input_tensors_needed and input_tensor_as_shapes_needed. Remove calls to call_cpp_shape_fn that pass input_tensors_needed or inptu_tensor_as_shapes_needed. Change: 139228180
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139224573
* Allow the Estimator default device function to merge devices in order to not ↵Gravatar Martin Wicke2016-11-15
| | | | | | break device assignments set by users. Change: 139224359
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139223923
* RGB Uint8 arrays for canvas label fill and stroke styles. Allows us to changeGravatar Charles Nicholson2016-11-15
| | | | | the label coloring for selected / hover / neighbor points, coming soon. Change: 139223883
* Always fully initialize PartitionOpts.Gravatar Asim Shankar2016-11-15
| | | | | Caught by a MemorySanitizer test failure. Change: 139222793
* adding docstrings to document which tensorflow functions have different ↵Gravatar Olivia Nordquist2016-11-15
| | | | | | names than the corresponding numpy function Change: 139221107
* Change constant_value for rank to return a scalar ndarray insteadGravatar A. Unique TensorFlower2016-11-15
| | | | | of a numpy array. Change: 139220642
* Update generated Python Op docs.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139218176
* Add tf-benchmark test for DNNLinearCombinedClassifier.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139217997
* tfdbg: exclude core/debug from iOS makefile buildGravatar Shanqing Cai2016-11-15
| | | | | in preparation for gRPC debug ops Change: 139216871
* Benchmark fused convolution ops, and use multi-threaded EigenTensor for conv2dGravatar Pete Warden2016-11-15
| | | | Change: 139215742
* Expose tf.reverse_v2 which takes indices rather than a bool tensor.Gravatar Andrew Selle2016-11-15
| | | | | | Also Implemented ReverseV2's gradient in C++ Up the version to indicate the new op is ready to be used. Change: 139215566
* Refactor DNNClassifier to use head.Gravatar A. Unique TensorFlower2016-11-15
| | | | Change: 139210204
* Switch Windows CI for CMake to use manually installed SWIG.Gravatar Derek Murray2016-11-15
| | | | | | | The choco-installed version of SWIG did not include the fix to https://github.com/swig/swig/issues/567, which caused some Python tests to fail on Windows. Change: 139206577
* Embedding Projector: Add google analytics for internal google use only.Gravatar Dan Smilkov2016-11-15
| | | | Change: 139202892
* C API: Do not take ownership of the TF_TensorsGravatar Asim Shankar2016-11-15
| | | | | | | | | | | | | | | Prior to this change, TF_*Run, TF_SetAttrTensor and TF_SetAttrTensorList took ownership of the TF_Tensor*s of the feeds. This can make performance client languages bothersome when the same Tensor is repeatedly fed into multiple session executions as the memory for the feed tensor would need to be re-allocated and filled in every time. With this change, these functions no longer take ownership of the TF_Tensor*. The changes to the Go API implementation reflect the claimed benefits. (Another step towards #10) Change: 139169388
* Update ops-related pbtxt files.Gravatar A. Unique TensorFlower2016-11-14
| | | | Change: 139163917
* Bring TensorArray into a new ref-free world.Gravatar Eugene Brevdo2016-11-14
| | | | Change: 139161477
* Seal tf.logging's interface.Gravatar Patrick Nguyen2016-11-14
| | | | Change: 139151696
* Generalized ExecutorImpl::InferAllocAttr to handle more types of accelerators.Gravatar Benoit Steiner2016-11-14
| | | | Change: 139149048