Code

Multi label models#67

Code

Merged

RichJackson merged 11 commits intomainAstraZeneca/KAZU:mainfrom multi-label-modelsAstraZeneca/KAZU:multi-label-modelsCopy head branch name to clipboardSep 18, 2024

+247-479

Commits

Commits on Sep 18, 2024

• Simplified Spanfinder:

  Show description for 708595a
  RichJackson

  committed
  708595a

• Remvoed unused IS_SUBSPAN constant

  RichJackson

  committed
  e892723

• TransformersModelForTokenClassificationNerStep now supports multilabel classification models.

  Show description for 20732bc
  RichJackson

  committed
  20732bc

• updated tests for TokenizedWordProcessor

  RichJackson

  committed
  6842ec0

• updated config for TransformersModelForTokenClassificationNerStep. Removed obsolete multilabel.yaml

  RichJackson

  committed
  86fedab

• fixed bug in HFDataset where all items were given the input_ids key, instead of their proper name

  RichJackson

  committed
  d34b662

• corrected input keys for default bert model

  RichJackson

  committed
  0570372

• removed redundant test

  RichJackson

  committed
  f85292e

• changed device param to string, with 'cpu' as default

  RichJackson

  committed
  cf42189

• fixed mixed up doc strings

  RichJackson

  committed
  6d46388

• bunch of nit picks on tokenized_word_processor.py

  Show description for b65fc0c
  

  RichJackson

  and

  Szymon Palucha

  committed
  b65fc0c