Commits

Branch selector

 master

User selector

qlzh727

Datepicker

All time

Commit History

Commits on Jun 14, 2019

• Update LSTM/GRU to have more strict test in v2 for graph rewriting.

  Show description for 8c8f573
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  8c8f573

  Copy full SHA for 8c8f573

Commits on Jun 12, 2019

• Disable flaky tests in OSS.

  Show description for 67d2bb8
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  67d2bb8

  Copy full SHA for 67d2bb8

• Automated rollback of commit 4b8a9fadd34f0bccde4b4f9666995a97cb02d9ff

  Show description for ca3b463
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  ca3b463

  Copy full SHA for ca3b463

• Internal change.

  Show description for cb9e43e
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  cb9e43e

  Copy full SHA for cb9e43e

Commits on Jun 11, 2019

• Revert the change to benchmark test, which was for local debugging.

  Show description for ed891e3
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  ed891e3

  Copy full SHA for ed891e3

Commits on Jun 10, 2019

• Fix the RNN backend swapping issue, and adding new unit test for model.eval.

  Show description for d837969
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  d837969

  Copy full SHA for d837969

• Internal change.

  Show description for e245b5b
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  e245b5b

  Copy full SHA for e245b5b

• Partially fix the function inlining and performance regression for LSTM/GRU.

  Show description for b2dfb9b
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  b2dfb9b

  Copy full SHA for b2dfb9b

• Internal change.

  Show description for a3fabf6
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  a3fabf6

  Copy full SHA for a3fabf6

Commits on Jun 7, 2019

• Automated rollback of commit e691be7814bcd7065950ec940456a4c9d8991645

  Show description for 09d4788
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  09d4788

  Copy full SHA for 09d4788

Commits on Jun 6, 2019

• Fix an important performance regression for LSTM and GRU in tf 2.0

  Show description for e691be7
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  e691be7

  Copy full SHA for e691be7

Commits on Jun 4, 2019

• Fix stateful RNN with recurrent_dropout in 2.0.

  Show description for 6a6e8c2
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  6a6e8c2

  Copy full SHA for 6a6e8c2

• Reset dropout masks for LSTM/GRU in v2 implementation.

  Show description for 180f28a
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  180f28a

  Copy full SHA for 180f28a

Commits on Jun 1, 2019

• Fix serial/deserialization issue for rnn wrapper.

  Show description for e62dc43
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  e62dc43

  Copy full SHA for e62dc43

Commits on May 30, 2019

• Add compute_output_signature for base_layer.

  Show description for 28cc95f
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  28cc95f

  Copy full SHA for 28cc95f

Commits on May 29, 2019

• Automated rollback of commit 6fcf520ae0935d96250e26ccbe89ca4bf20d6174

  Show description for 5d74b48
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  5d74b48

  Copy full SHA for 5d74b48

• Refactor training_distributed to reduce the duplication.

  Show description for c362fa7
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  c362fa7

  Copy full SHA for c362fa7

Commits on May 28, 2019

• Internal cleanup.

  Show description for c47ec54
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  c47ec54

  Copy full SHA for c47ec54

• Fix Bidirectional wrapper issue for eager/np inputs.

  Show description for 2a8f9b1
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  2a8f9b1

  Copy full SHA for 2a8f9b1

Commits on May 23, 2019

• Refactor the training logic based on the input data type.

  Show description for e211a81
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  e211a81

  Copy full SHA for e211a81

• Internal change.

  Show description for b6f1844
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  b6f1844

  Copy full SHA for b6f1844

Commits on May 21, 2019

• Fix LSTM state issue for 2.7.

  Show description for 1bc7dd2
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  1bc7dd2

  Copy full SHA for 1bc7dd2

Commits on May 15, 2019

• Fix customized cell handling for Bidirectional wrapper.

  Show description for 0667725
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  0667725

  Copy full SHA for 0667725

Commits on May 14, 2019

• Fix stateful RNN with stack cell that has multiple states.

  Show description for 1225055
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  1225055

  Copy full SHA for 1225055

Commits on May 13, 2019

• RNN test cleanup.

  Show description for f587874
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  f587874

  Copy full SHA for f587874

• Consolidate sample weights and mode into TrainingEndpoint.

  Show description for 26a074b
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  26a074b

  Copy full SHA for 26a074b

• Consolidating more model output related attributes in the training loop.

  Show description for 15052fa
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  15052fa

  Copy full SHA for 15052fa

Commits on May 8, 2019

• Update implementation_selector to swallow the error during function load.

  Show description for 24f01a8
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  24f01a8

  Copy full SHA for 24f01a8

Commits on May 7, 2019

• Refactoring the keras compile() by consolidating internal fields.

  Show description for e905b8b
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  e905b8b

  Copy full SHA for e905b8b

Commits on May 6, 2019

• Fix mask function serialization for lambda layer.

  Show description for 2707117
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  2707117

  Copy full SHA for 2707117

Commits on May 3, 2019

• Refactor for Keras model attributes and model.compile().

  Show description for 0a782f1
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  0a782f1

  Copy full SHA for 0a782f1

Commits on May 2, 2019

• Allow user to specify backward layer for bidirectional RNN.

  Show description for fe5e1f3
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  fe5e1f3

  Copy full SHA for fe5e1f3

• Automated rollback of commit 8442c53ecaeff0544e3b6c07a77ba03f6f9ec63f

  Show description for 6e28586
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  6e28586

  Copy full SHA for 6e28586

• Allow user to specify backward layer for bidirectional RNN.

  Show description for 8442c53
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  8442c53

  Copy full SHA for 8442c53

Commits on Apr 29, 2019

• Add unit test to verify the masking support of lambda layer.

  Show description for e5a0e8e
  

  qlzh727

  authored and

  tensorflower-gardener

  committed
  e5a0e8e

  Copy full SHA for e5a0e8e

Pagination

Previous

Next