982549ea3423df4270ff154e5c764beb43d472da |
|
26-Jan-2018 |
Rasmus Munk Larsen <rmlarsen@google.com> |
Branch 183429339 (#16469) * Change `reduce_logsumexp` to internally use `reshape` rather than `squeeze` since the latter requires the `axis` arg to be a Python `list`. PiperOrigin-RevId: 183396533 * Kernel utils to support broadcast add and mul. PiperOrigin-RevId: 183397494 * Updating sparsify_gather. PiperOrigin-RevId: 183402917 * [tf.data] Move slow-path-related code into the slow path in IteratorHandleOp::Compute(). This slightly reduces the amount of work performed when an iterator is accessed (after the first access), and potentially reduces contention if concurrent steps are accessing the same iterator. PiperOrigin-RevId: 183406221 * Cleanup: Ran clang-format on all *.{cc,h} in under grappler. PiperOrigin-RevId: 183406440 * Increase shard count of //third_party/tensorflow/python:nn_batchnorm_test to avoid timeouts When run under asan, the test runs for about 5 minutes, and sometimes longer, causing frequent timeouts. This change increases the shard count of the test to 4, which brings the run time of the longest running shard under asan to about 2 minutes. PiperOrigin-RevId: 183414888 * Add available choices to toco flags and fix minor formatting issues. PiperOrigin-RevId: 183415713 * Performance improvements to some GPU code to use shared locks instead of unique locks for some hotspot cases. PiperOrigin-RevId: 183418559 * [XLA] Improve error message for bad slices. PiperOrigin-RevId: 183420038 * Fix py3 build rules for all py tests under py2tf. PiperOrigin-RevId: 183422144 * Fix bug with Operation._control_inputs setter. PiperOrigin-RevId: 183422192 * Make softmax_op_test.py work with C API enabled. PiperOrigin-RevId: 183422829 * Cleanup: Ran clang-format on all *.{cc,h} files in tensorflow/core/kernels. PiperOrigin-RevId: 183423961 * Fix the documentation for the dense layer for how rank > 2 inputs are handled. PiperOrigin-RevId: 183425868 * Cleanup: Ran clang-format on all *.{cc,h} in tensorflow/core/ops. PiperOrigin-RevId: 183429339
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
f8347ceebbad0e06552633fcdf8e63f52246ba62 |
|
24-Jan-2018 |
Sanjoy Das <sanjoy@google.com> |
Remove THIRD_PARTY_ from #include guards They don't make sense in the open source repository. PiperOrigin-RevId: 183140889
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
c8b59c046895fa5b6d79f73e0b5817330fcfbfc1 |
|
02-Jun-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
Update copyright for 3p/tf/core. Change: 123900938
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
5a130818175ae5f9551012ecdb0f38854c1a72a3 |
|
29-Mar-2016 |
Benoit Steiner <benoit.steiner.goog@gmail.com> |
Leverage index list to further speedup the computation of the convolution gradients by 1 to 10% depending on the size of the convolution kernel. Change: 118505660
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
6a073b392fd270bb1b1034b2e096b8b7b5b9f83e |
|
29-Mar-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
First pass at enabling non-square strides for convolutions: - Change the stride, and in_stride template arguments of Eigen SpatialConvolution, SpatialConvolutionBackwardInput, and SpatialConvolutionBackwardKernel to row_stride, col_stride, row_in_stride, col_in_stride. - Change tensorflow kernels to pass the additional stride parameters. - Rationalize the place where we swap row/col: swap just before calling Eigen. This just enables the plumbing. Non-square strides are still forbidden in the ops. Change: 118484322
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
f4acbbf24904f5ffd8ad487461095bcde0886983 |
|
29-Mar-2016 |
Benoit Steiner <benoit.steiner.goog@gmail.com> |
Further improved the performance of the backard pass of the convolution Change: 118414827
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
1e1beefce2f40dc13f3374fdff4a83b63196d070 |
|
29-Mar-2016 |
A. Unique TensorFlower <nobody@tensorflow.org> |
250% GPU speed up of the convolution gradient computation wrt the weights for Eigen. Change: 118408303
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
da767e1079dfc683b126c0ba69ed6ced925c3300 |
|
25-Mar-2016 |
Benoit Steiner <benoit.steiner.goog@gmail.com> |
Sped up the computation of the gradients of the convolution layer wrt the inputs by 30% on GPU. Change: 118221540
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
7323103c1c067ba776a6acf25af82ffbec1290ed |
|
23-Mar-2016 |
Benoit Steiner <benoit.steiner.goog@gmail.com> |
Move the NeuralNetwork code out of third_party/eigen3 and into tensorflow/core/kernel. Change: 117941211
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
4671953808f87daa6c8355bdd961242290aab80f |
|
19-Mar-2016 |
Vijay Vasudevan <vrv@google.com> |
Re-rollback of "TensorFlow: move eigen some NN code from our third_party/eigen3 copy to being part of TF, add tests." Change: 117608627
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
8202d0958e6894f91ff7a397a00a192fe306eb95 |
|
18-Mar-2016 |
Vijay Vasudevan <vrv@google.com> |
Rollforward of "TensorFlow: move eigen some NN code from our third_party/eigen3 copy to being part of TF, add tests." Change: 117587217
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
200fa5fb0be1523f828014ae8f3ca28099cc942a |
|
18-Mar-2016 |
Vijay Vasudevan <vrv@google.com> |
Rollback of "TensorFlow: move eigen some NN code from our third_party/eigen3 copy to being part of TF, add tests." Change: 117519243
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|
50edc4e878c076c44fecc847af110a19b171eb63 |
|
18-Mar-2016 |
Vijay Vasudevan <vrv@google.com> |
TensorFlow: move eigen some NN code from our third_party/eigen3 copy to being part of TF, add tests. Change: 117509710
/external/tensorflow/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h
|