Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Oct 20, 2020

• Bump version to 2.13.0

  guillaumekln

  committed
  ff2f5a9

  Copy full SHA for ff2f5a9

• Update CHANGELOG.md

  guillaumekln

  authored
  c0e34e1

  Copy full SHA for c0e34e1

• Ignore evaluation and checkpoint when computing training throughput (#731)

  Show description for 86147ec
  guillaumekln

  authored
  86147ec

  Copy full SHA for 86147ec

• Cleanup asset prefix generation and usage (#730)

  Show description for e232faa
  guillaumekln

  authored
  e232faa

  Copy full SHA for e232faa

Commits on Oct 19, 2020

• Make error more explicit when the input layer is not initialized (#729)

  guillaumekln

  authored
  492f6fc

  Copy full SHA for 492f6fc

• Expose function to create TensorFlow lookup tables (#728)

  guillaumekln

  authored
  b62310e

  Copy full SHA for b62310e

Commits on Oct 16, 2020

• Use single method to create all optimizer weights (#727)

  guillaumekln

  authored
  821eccc

  Copy full SHA for 821eccc

• Update CHANGELOG.md

  guillaumekln

  authored
  b42a1b0

  Copy full SHA for b42a1b0

• Allow RaggedTensor as encoder inputs (#726)

  guillaumekln

  authored
  c8f915d

  Copy full SHA for c8f915d

Commits on Oct 15, 2020

• Fix typo in vocabulary.md

  guillaumekln

  authored
  0a4b2f0

  Copy full SHA for 0a4b2f0

• Get version from the package when building docs (#725)

  guillaumekln

  authored
  866b0dd

  Copy full SHA for 866b0dd

Commits on Oct 14, 2020

• Return a summary of the training from the train method (#724)

  Show description for 34a5f52
  guillaumekln

  authored
  34a5f52

  Copy full SHA for 34a5f52

Commits on Oct 9, 2020

• Fix missing closing parenthesis in example

  guillaumekln

  authored
  36524df

  Copy full SHA for 36524df

• Fix small typo in README

  guillaumekln

  authored
  e488366

  Copy full SHA for e488366

• Add OpenNMT-tf version in the training logs (#722)

  Show description for 1c9fca8
  guillaumekln

  authored
  1c9fca8

  Copy full SHA for 1c9fca8

Commits on Sep 24, 2020

• Ignore first iterations in throughput reporting (#718)

  guillaumekln

  authored
  d185690

  Copy full SHA for d185690

Commits on Sep 23, 2020

• Remove comment that describes a scenario not supported by the class

  guillaumekln

  committed
  6f5a892

  Copy full SHA for 6f5a892

• Update PyPI classifiers

  guillaumekln

  committed
  6286587

  Copy full SHA for 6286587

Commits on Sep 18, 2020

• Raise errors when starting the training but end conditions are met (#716)

  Show description for 21df1c7
  guillaumekln

  authored
  21df1c7

  Copy full SHA for 21df1c7

Commits on Sep 16, 2020

• Bump version to 2.12.1

  guillaumekln

  committed
  31a897e

  Copy full SHA for 31a897e

• Update sacrebleu to 1.4.14 to fix Python 3.5 compatibility (#715)

  guillaumekln

  authored
  1ba1871

  Copy full SHA for 1ba1871

Commits on Sep 7, 2020

• Remove TODO

  Show description for 032918c
  guillaumekln

  committed
  032918c

  Copy full SHA for 032918c

Commits on Aug 31, 2020

• Update serving examples based on the latest model export (#712)

  Show description for ee8e1dd
  guillaumekln

  authored
  ee8e1dd

  Copy full SHA for ee8e1dd

• Bump version to 2.12.0

  guillaumekln

  committed
  df74060

  Copy full SHA for df74060

• Update CHANGELOG.md

  guillaumekln

  authored
  48953da

  Copy full SHA for 48953da

Commits on Aug 7, 2020

• Make DecodingStrategy methods public (#711)

  guillaumekln

  authored
  fe94c1c

  Copy full SHA for fe94c1c

Commits on Aug 6, 2020

• Update CONTRIBUTING.md

  guillaumekln

  authored
  9c702fc

  Copy full SHA for 9c702fc

• Check that training logs do not raise exceptions

  guillaumekln

  committed
  4a16d9a

  Copy full SHA for 4a16d9a

• Run keep_for_training in inputter tests

  guillaumekln

  committed
  84fee0c

  Copy full SHA for 84fee0c

• Check that Inputter.make_features is runnable in eager mode

  guillaumekln

  committed
  a8af05f

  Copy full SHA for a8af05f

• Check that inference time logging does not raise exceptions

  guillaumekln

  committed
  221acff

  Copy full SHA for 221acff

• Ignore coverage for abstract methods

  guillaumekln

  committed
  8de3b03

  Copy full SHA for 8de3b03

• Add minimal tests for WER and TER scorers

  guillaumekln

  committed
  5bd82e3

  Copy full SHA for 5bd82e3

• Remove unused function_on_next decorator

  guillaumekln

  committed
  ac67097

  Copy full SHA for ac67097

• Update to TensorFlow 2.3 and TensorFlow Addons 0.11 (#710)

  Show description for 56672c8
  guillaumekln

  authored
  56672c8

  Copy full SHA for 56672c8

Pagination

Previous

Next