Commits

Branch selector

 master

User selector

All users

All time

Commit History

Commits on Dec 17, 2018

• [XLA] Improve runtime of MakeInstructionPostOrder by reserving the flat set

  Show description for feaaa82
  

  blakehechtman

  authored and

  tensorflower-gardener

  committed
  feaaa82

  Copy full SHA for feaaa82

Commits on Dec 16, 2018

• compat: Update forward compatibility horizon to 2018-12-16

  Show description for ebe98d9
  tensorflower-gardener

  committed
  ebe98d9

  Copy full SHA for ebe98d9

• Fix path name in the non gpu case.

  Show description for 39d9c9e
  

  aselle

  authored and

  tensorflower-gardener

  committed
  39d9c9e

  Copy full SHA for 39d9c9e

Commits on Dec 15, 2018

• compat: Update forward compatibility horizon to 2018-12-15

  Show description for b37a329
  tensorflower-gardener

  committed
  b37a329

  Copy full SHA for b37a329

• Return a clear error message when outside compilation has int64 input/output.

  Show description for 65a0a3a
  

  er91

  authored and

  tensorflower-gardener

  committed
  65a0a3a

  Copy full SHA for 65a0a3a

• Merge status values when aggregating status from executor or multi-graph runner.

  Show description for 65470e7
  

  rjpower

  authored and

  tensorflower-gardener

  committed
  65470e7

  Copy full SHA for 65470e7

• This CL builds the capability of traversing the python graph to determine the original input nodes. We also make the error message more readable by removing common path prefix from stack traces

  Show description for 5138caa
  tensorflower-gardener

  committed
  5138caa

  Copy full SHA for 5138caa

• Extended experimental API to allow the configuration of the number of CPU devices.

  Show description for c7c0a76
  

  Mingsheng Hong

  authored and

  tensorflower-gardener

  committed
  c7c0a76

  Copy full SHA for c7c0a76

• Use logits directly when using sparse categorical cross entropy loss function

  Show description for 2a0561d
  tensorflower-gardener

  committed
  2a0561d

  Copy full SHA for 2a0561d

• [XLA] Instead of ever recomputing reachability during fusion, use the original

  Show description for aab3185
  

  blakehechtman

  authored and

  tensorflower-gardener

  committed
  aab3185

  Copy full SHA for aab3185

• Merge pull request #24382 from tfboyd:docker_cuda10

  Show description for 84d4bfc
  tensorflower-gardener

  committed
  84d4bfc

  Copy full SHA for 84d4bfc

• Move Keras to SavedModel code and some SavedModel utils into core.

  Show description for 212d8cd
  

  k-w-w

  authored and

  tensorflower-gardener

  committed
  212d8cd

  Copy full SHA for 212d8cd

• Disable automatic dependency tracking for the legacy tf.layers.Layer class.

  Show description for 98c0deb
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  98c0deb

  Copy full SHA for 98c0deb

• Automated rollback of commit d19a8c92a23db3cc5c2ae74a8443b16d6d48eb13

  Show description for c152e8c
  

  gmagogsfm

  authored and

  tensorflower-gardener

  committed
  c152e8c

  Copy full SHA for c152e8c

• Add UnifiedGRU as the new GRU implementation for tf2.0.

  Show description for a8009f4
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  a8009f4

  Copy full SHA for a8009f4

Commits on Dec 14, 2018

• Add zip and enumerate to the builtins list.

  Show description for d5d3416
  

  Dan Moldovan

  authored and

  tensorflower-gardener

  committed
  d5d3416

  Copy full SHA for d5d3416

• Automated rollback of commit e9e534f3239d7cb7e2a815b49c1c3520d9566d70

  Show description for 720cb66
  

  alextp

  authored and

  tensorflower-gardener

  committed
  720cb66

  Copy full SHA for 720cb66

• Create v2 ModeKeys class that's shared among TensorFlow (Keras + Estimator). It will be exported as tf.estimator.ModeKeys in a follow-up change.

  Show description for 8bcc801
  

  rchao

  authored and

  tensorflower-gardener

  committed
  8bcc801

  Copy full SHA for 8bcc801

• Automated rollback of commit 306cf4b2834cca1f1fc1fc58ba16c3248516dfd9

  Show description for e2acc3d
  tensorflower-gardener

  committed
  e2acc3d

  Copy full SHA for e2acc3d

• Add Callbacks hooks for `evaluate` and `predict`.

  Show description for abe1c5f
  tensorflower-gardener

  committed
  abe1c5f

  Copy full SHA for abe1c5f

• More weakreffing in function

  Show description for 62e61f4
  

  allenlavoie

  authored and

  tensorflower-gardener

  committed
  62e61f4

  Copy full SHA for 62e61f4

• Remove one last(-ish) reference cycle from tf.function

  Show description for 777eaf2
  

  allenlavoie

  authored and

  tensorflower-gardener

  committed
  777eaf2

  Copy full SHA for 777eaf2

• Temporarily disable an EXPECT_GE in port_test. This EXPECT_GE currently fails on

  Show description for 55bc1b5
  

  fdxmw

  authored and

  tensorflower-gardener

  committed
  55bc1b5

  Copy full SHA for 55bc1b5

• Fix error message when bucketized_column is missing boundaries.

  Show description for 6634bc8
  tensorflower-gardener

  committed
  6634bc8

  Copy full SHA for 6634bc8

• Deprecated Variable.initialized_value and Variable.load in eager

  Show description for 11ea11d
  

  superbobry

  authored and

  tensorflower-gardener

  committed
  11ea11d

  Copy full SHA for 11ea11d

• Automatically clean up tf.function reference cycles

  Show description for 5926ac2
  

  allenlavoie

  authored and

  tensorflower-gardener

  committed
  5926ac2

  Copy full SHA for 5926ac2

• Fix broken benchmarks

  Show description for ece6ba3
  

  Rachel Lim

  authored and

  tensorflower-gardener

  committed
  ece6ba3

  Copy full SHA for ece6ba3

• Merge pull request #24353 from trevor-m:tmorris_tftrt_determinism

  Show description for f750433
  tensorflower-gardener

  committed
  f750433

  Copy full SHA for f750433

• Upgrade to CUDA 10

  tfboyd

  committed
  29b8d49

  Copy full SHA for 29b8d49

• Make the default exponent a float so that when using override_from_dict() or set() float values aren't rejected.

  Show description for d52e7e1
  tensorflower-gardener

  committed
  d52e7e1

  Copy full SHA for d52e7e1

• Moved `Variable.constraint` to Keras.

  Show description for e9e534f
  

  superbobry

  authored and

  tensorflower-gardener

  committed
  e9e534f

  Copy full SHA for e9e534f

• Further cleanup for RunInputs for fuzzers.

  Show description for 06c3180
  

  mihaimaruseac

  authored and

  tensorflower-gardener

  committed
  06c3180

  Copy full SHA for 06c3180

• Adds numerical correctness tests for all Keras modes and model types

  Show description for e8d4a3d
  tensorflower-gardener

  committed
  e8d4a3d

  Copy full SHA for e8d4a3d

• Update cuda 10 image with TensorRT.

  Show description for 690bd62
  tensorflower-gardener

  committed
  690bd62

  Copy full SHA for 690bd62

• Fix Keras tests by using TF optimizers in tests with run_eagerly enabled

  Show description for 89b11a3
  

  fchollet

  authored and

  tensorflower-gardener

  committed
  89b11a3

  Copy full SHA for 89b11a3