Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Dec 24, 2018

• Disable grappler remapping, to avoid introducing unsupported ops. Also: avoid removing the input nodes of a graph.

  Show description for 9f23c56
  tensorflower-gardener

  committed
  9f23c56

  Copy full SHA for 9f23c56

Commits on Dec 23, 2018

• [XLA] Allow non-monotonic batch dimensions, they are supported by the

  Show description for 6e304fa
  

  blakehechtman

  authored and

  tensorflower-gardener

  committed
  6e304fa

  Copy full SHA for 6e304fa

• Adding V2 categorical hinge loss and metric.

  Show description for 1853b08
  

  pavithrasv

  authored and

  tensorflower-gardener

  committed
  1853b08

  Copy full SHA for 1853b08

• compat: Update forward compatibility horizon to 2018-12-23

  Show description for ad6d0cb
  tensorflower-gardener

  committed
  ad6d0cb

  Copy full SHA for ad6d0cb

• Re-prefetch the dataset for V1Adapater

  Show description for 0a708f8
  

  Jianwei Xie

  authored and

  tensorflower-gardener

  committed
  0a708f8

  Copy full SHA for 0a708f8

• Disable cost tracking for async kernels.

  Show description for 3785cff
  

  rjpower

  authored and

  tensorflower-gardener

  committed
  3785cff

  Copy full SHA for 3785cff

Commits on Dec 22, 2018

• Enforce the contract of and improve the documentation for

  Show description for b27d50f
  tensorflower-gardener

  committed
  b27d50f

  Copy full SHA for b27d50f

• Replacing convert_stack(op._traceback) with op.traceback which internally calls the same function convert_stack(...). In this way, we don't have to reference internal property of the Operation object.

  Show description for cd8c6c9
  tensorflower-gardener

  committed
  cd8c6c9

  Copy full SHA for cd8c6c9

• compat: Update forward compatibility horizon to 2018-12-22

  Show description for bc81706
  tensorflower-gardener

  committed
  bc81706

  Copy full SHA for bc81706

• Allow dot to have mutiple contracting dimensions and non contracting dimensions.

  Show description for fa76858
  

  blakehechtman

  authored and

  tensorflower-gardener

  committed
  fa76858

  Copy full SHA for fa76858

• UnidirectionSequenceRNN hybrid int8 support.

  Show description for b13f1e5
  

  Suharsh Sivakumar

  authored and

  tensorflower-gardener

  committed
  b13f1e5

  Copy full SHA for b13f1e5

• Automated rollback of commit 20796e6bebb3a411655157f7ee7201af58a7d2fe

  Show description for 2d1964e
  

  annarev

  authored and

  tensorflower-gardener

  committed
  2d1964e

  Copy full SHA for 2d1964e

• [Grappler] Use input instead of node name when swapping control dependency in ConstantPushDown of ConstantFolding

  Show description for 6f81b31
  

  andyly

  authored and

  tensorflower-gardener

  committed
  6f81b31

  Copy full SHA for 6f81b31

• Uni Sequence LSTM int8 hybrid support.

  Show description for f080ab4
  

  Suharsh Sivakumar

  authored and

  tensorflower-gardener

  committed
  f080ab4

  Copy full SHA for f080ab4

• [XLA] Avoid UB in HloEvaluator by doing signed + and - using unsigned integers.

  Show description for 2be3404
  

  Justin Lebar

  authored and

  tensorflower-gardener

  committed
  2be3404

  Copy full SHA for 2be3404

• Turn on MKL-DNN contraction kernels by default. To disable them, build with --define=tensorflow_mkldnn_contraction_kernel=0.

  Show description for 20796e6
  

  penpornk

  authored and

  tensorflower-gardener

  committed
  20796e6

  Copy full SHA for 20796e6

• Adding V2 squared hinge loss and metric.

  Show description for 4132f1e
  

  pavithrasv

  authored and

  tensorflower-gardener

  committed
  4132f1e

  Copy full SHA for 4132f1e

• Renaming the tensor_checkpoint function as tensor_tracepoint.

  Show description for adf5654
  tensorflower-gardener

  committed
  adf5654

  Copy full SHA for adf5654

• Internal change.

  Show description for d5bd493
  tensorflower-gardener

  committed
  d5bd493

  Copy full SHA for d5bd493

• [XLA] Support kComplex in HloEvaluator.

  Show description for acb2cf1
  

  kayzhu

  authored and

  tensorflower-gardener

  committed
  acb2cf1

  Copy full SHA for acb2cf1

• [XLA] Make Shape <-> ShapeProto conversion include dynamic dimensions.

  Show description for d7cfff9
  

  congliuthu

  authored and

  tensorflower-gardener

  committed
  d7cfff9

  Copy full SHA for d7cfff9

• Dynamic expensive kernel tracking.

  Show description for 9a48681
  

  rjpower

  authored and

  tensorflower-gardener

  committed
  9a48681

  Copy full SHA for 9a48681

• [tf.data] Minor refactoring of the parallel interleave kernel.

  Show description for 0dd93b6
  

  jsimsa

  authored and

  tensorflower-gardener

  committed
  0dd93b6

  Copy full SHA for 0dd93b6

• Explicitly depend on keras_lib target in Estimator code.

  Show description for 7ca26f4
  

  Michael Case

  authored and

  tensorflower-gardener

  committed
  7ca26f4

  Copy full SHA for 7ca26f4

Commits on Dec 21, 2018

• Implement dynamic learning rate support for TPU embedding.

  Show description for 6f7d5cd
  tensorflower-gardener

  committed
  6f7d5cd

  Copy full SHA for 6f7d5cd

• Allow Keras Networks that are graph networks to accept nested inputs and

  Show description for c2f1888
  tensorflower-gardener

  committed
  c2f1888

  Copy full SHA for c2f1888

• Fix lstm and svdf tests

  Show description for fa1406e
  tensorflower-gardener

  committed
  fa1406e

  Copy full SHA for fa1406e

• Go: Update generated wrapper functions for TensorFlow ops.

  Show description for ce33494
  tensorflower-gardener

  committed
  ce33494

  Copy full SHA for ce33494

• [XLA] Return an error for xla::Cholesky() for complex types.

  Show description for fa9fe7d
  

  hawkinsp

  authored and

  tensorflower-gardener

  committed
  fa9fe7d

  Copy full SHA for fa9fe7d

• Split V1 tests out of cudnn_recurrent_test.py into cudnn_recurrent_v1_test.py

  Show description for 44c8cd2
  tensorflower-gardener

  committed
  44c8cd2

  Copy full SHA for 44c8cd2

• Adding V2 Hinge loss and metric

  Show description for eb1f8c4
  

  pavithrasv

  authored and

  tensorflower-gardener

  committed
  eb1f8c4

  Copy full SHA for eb1f8c4

• Use GRPC_ARG_MAX_RECONNECT_BACKOFF_MS instead of "grpc.testing.fixed_reconnect_backoff_ms" to restrict the backoff to 1s

  Show description for a752aea
  tensorflower-gardener

  committed
  a752aea

  Copy full SHA for a752aea

• Update ops-related pbtxt files.

  Show description for d1ef850
  tensorflower-gardener

  committed
  d1ef850

  Copy full SHA for d1ef850

• Add _REF() registration to a few ops.

  Show description for 5ff9d98
  

  liyunlu0618

  authored and

  tensorflower-gardener

  committed
  5ff9d98

  Copy full SHA for 5ff9d98

• Test case cleanup.

  Show description for a15adad
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  a15adad

  Copy full SHA for a15adad

Pagination

Previous

Next