Skip to content
This repository has been archived by the owner on Mar 3, 2024. It is now read-only.

Latest commit

 

History

History
83 lines (44 loc) · 1.03 KB

CHANGELOG.md

File metadata and controls

83 lines (44 loc) · 1.03 KB

Changelog

[Unreleased]

[0.85.0] - 2020-07-09

Fixed

  • Compatible with Keras 2.4.3

[0.82.0] - 2020-06-02

Removed

  • Adapter

[0.78.0] - 2019-09-17

Fixed

  • Compatible with Keras 2.3.0

[0.70.0] - 2019-07-16

Added

  • Try to find the indices of tokens in the original text.

[0.69.0] - 2019-07-16

Added

[0.60.0] - 2019-06-10

Added

  • trainable can be a list of prefixes of layer names

[0.58.0] - 2019-06-10

Fixed

  • Use math_ops for tensorflow backend
  • Assign names to variables in warmup optimizer

[0.56.0] - 2019-06-04

Changed

  • Docs about training and trainable

Fixed

  • Missing trainable=False when training=True

[0.54.0] - 2019-05-29

Added

  • Support eager mode with tensorflow backend

[0.43.0] - 2019-05-12

Added

  • Support tf.keras

[0.40.0] - 2019-04-29

Added

  • Warmup optimizer

[Older Versions]

Added

  • BERT implementation
  • Load official model
  • Tokenizer
  • Demos