Commits

Branch selector

 master

User selector

tensorflower-gardener

Datepicker

All time

Commit History

Commits on Oct 13, 2017

• Fix a dtype

  Show description for 943feb0
  tensorflower-gardener

  committed
  943feb0

  Copy full SHA for 943feb0

• Removing custom CROSSTOOL and BUILD file for TF cpu builds as they are redundant now that bazel-toolchains repo is live

  Show description for bf84210
  tensorflower-gardener

  committed
  bf84210

  Copy full SHA for bf84210

• Direct users of the old KMeans estimator to use the new one that's built from the core Estimator API.

  Show description for aea2a31
  tensorflower-gardener

  committed
  aea2a31

  Copy full SHA for aea2a31

• Internal change.

  Show description for 7061d2a
  tensorflower-gardener

  committed
  7061d2a

  Copy full SHA for 7061d2a

• Avoid cache thrashing in CTC beam search

  Show description for 4503f46
  tensorflower-gardener

  committed
  4503f46

  Copy full SHA for 4503f46

• Small improvement of verification heuristics.

  Show description for 7e31a19
  tensorflower-gardener

  committed
  7e31a19

  Copy full SHA for 7e31a19

• Extend the ExecuteParallel service interface to allow multiple devices per computation.

  Show description for 2645045
  tensorflower-gardener

  committed
  2645045

  Copy full SHA for 2645045

• Add "clear_output_shapes" option to FoldConstants transformer in

  Show description for a06b378
  tensorflower-gardener

  committed
  a06b378

  Copy full SHA for a06b378

• Add streaming_false_{negative,positive}_rate and streaming_false_{negative,positive}_rate_at_thresholds.

  Show description for 7156831
  tensorflower-gardener

  committed
  7156831

  Copy full SHA for 7156831

• Removes brittleness from some tests in dnn_test.

  Show description for 215aca5
  tensorflower-gardener

  committed
  215aca5

  Copy full SHA for 215aca5

• Switch to nest.flatten() in tf.layers.Layer to allow dicts and

  Show description for cb7dca2
  tensorflower-gardener

  committed
  cb7dca2

  Copy full SHA for cb7dca2

• [TF:XLA] Implement BitwiseAnd, BitwiseOr, and Invert operators.

  Show description for 915a8ac
  tensorflower-gardener

  committed
  915a8ac

  Copy full SHA for 915a8ac

• typo

  Show description for 4b17895
  tensorflower-gardener

  committed
  4b17895

  Copy full SHA for 4b17895

• BUILD cleanup in contrib/boosted_trees

  Show description for 6ef5428
  tensorflower-gardener

  committed
  6ef5428

  Copy full SHA for 6ef5428

Commits on Oct 12, 2017

• A few profiler improvements.

  Show description for 19708cc
  tensorflower-gardener

  committed
  19708cc

  Copy full SHA for 19708cc

• Puts the global step read within the global step name scope, for clean/organized graph displays in TensorBoard.

  Show description for 1002f97
  tensorflower-gardener

  committed
  1002f97

  Copy full SHA for 1002f97

• Replace tf.conj(tf.transpose(foo)) and tf.transpose(tf.conj(foo)) idioms with tf.linalg.adjoint(foo) or tf.transpose(foo, conjugate=True), and clean up a few places that can avoid explicit adjoints…

  Show description for bc52fbd
  tensorflower-gardener

  committed
  bc52fbd

  Copy full SHA for bc52fbd

• Make CheckNumerics on GPU asynchronous and remove the need for a hard device synchronization, which can have a significant performance impact.

  Show description for d6f817a
  tensorflower-gardener

  committed
  d6f817a

  Copy full SHA for d6f817a

• Require that work_element_count be positive in GetCudaLaunchConfig() rather than trying to launch kernels with an invalid config and leaving the stream in an error state. The latter is much harder …

  Show description for 9d111d9
  tensorflower-gardener

  committed
  9d111d9

  Copy full SHA for 9d111d9

• Add linux_arm64 and linux_armhf config_setting.

  Show description for abb6b8e
  tensorflower-gardener

  committed
  abb6b8e

  Copy full SHA for abb6b8e

• Use canonical method for not changing the scope, but changing scope options.

  Show description for f7b3229
  tensorflower-gardener

  committed
  f7b3229

  Copy full SHA for f7b3229

• Explictly raise an exception if sharding fails because the batch size is unknown.

  Show description for 1885344
  tensorflower-gardener

  committed
  1885344

  Copy full SHA for 1885344

• Update ops-related pbtxt files.

  Show description for 8c57896
  tensorflower-gardener

  committed
  8c57896

  Copy full SHA for 8c57896

• This creates a library under contrib for implementations of common metric learning losses,

  Show description for deb72df
  tensorflower-gardener

  committed
  deb72df

  Copy full SHA for deb72df

• eager: Expose tfe.run_test_in_graph_and_eager_modes decorator. Useful when writing tests of libraries.

  Show description for aa7532c
  tensorflower-gardener

  committed
  aa7532c

  Copy full SHA for aa7532c

• Add warm_starting_utils for Estimators.

  Show description for 76524d0
  tensorflower-gardener

  committed
  76524d0

  Copy full SHA for 76524d0

• Minimal support for running OpsTest on GPU, using CUDA unified memory.

  Show description for d244ffb
  tensorflower-gardener

  committed
  d244ffb

  Copy full SHA for d244ffb

• BUILD cleanup in contrib/boosted_trees/...

  Show description for 66b1615
  tensorflower-gardener

  committed
  66b1615

  Copy full SHA for 66b1615

• Optimized C++ and CUDA kernels for matrix_set_diag op. The new code is faster and more readable and avoids an issue with using the Eigen generator mechanism with GPUs on Windows.

  Show description for cec93f1
  tensorflower-gardener

  committed
  cec93f1

  Copy full SHA for cec93f1

• Go: Update generated wrapper functions for TensorFlow ops.

  Show description for e32292c
  tensorflower-gardener

  committed
  e32292c

  Copy full SHA for e32292c

• Update ops-related pbtxt files.

  Show description for e0e6fee
  tensorflower-gardener

  committed
  e0e6fee

  Copy full SHA for e0e6fee

• Improves error message when labels is None.

  Show description for 27a7e5c
  tensorflower-gardener

  committed
  27a7e5c

  Copy full SHA for 27a7e5c

• Optimize gradients for Mean

  Show description for 915e60d
  tensorflower-gardener

  committed
  915e60d

  Copy full SHA for 915e60d

• Implement NCHW_VECT_C support for tf.depth_to_space on GPU.

  Show description for 9b26ed7
  tensorflower-gardener

  committed
  9b26ed7

  Copy full SHA for 9b26ed7

• Update ops-related pbtxt files.

  Show description for c69b959
  tensorflower-gardener

  committed
  c69b959

  Copy full SHA for c69b959

Pagination

Previous

Next