eubinecto/idiomify

Issues

• Chronicles

  #4 opened 3 years ago by eubinecto
  1

• login with OpenAI token

  #28 opened 3 years ago by eubinecto
  2

• `main_infer.py`: don't split sentences

  #27 opened 3 years ago by eubinecto
  1

• remove the special tokens

  #26 opened 3 years ago by eubinecto
  0

• add password check

  #24 opened 3 years ago by eubinecto
  0

• Why does fine-tuning perform worse on the same data?

  #23 opened 3 years ago by eubinecto
  4

• fine-tune Davinci-002

  #18 opened 3 years ago by eubinecto
  9

• add a simple password check

  #22 opened 3 years ago by eubinecto
  1

• build `literal2idiomatic` dataset (at least three exemplar usages for each idiom)

  #21 opened 3 years ago by eubinecto
  3

• Choose the idioms to collect data for

  #20 opened 3 years ago by eubinecto
  4

• Implement `Idiomifier`class with OpenAI's GPT-3 API

  #16 opened 3 years ago by eubinecto
  3

• : Idiomfier as an NER tagger

  #15 opened 3 years ago by eubinecto
  1

• `entities:d-1-4`: define the entities

  #12 opened 3 years ago by eubinecto
  0

• `d-1-4` : Preprocess PIE dataset to build NER labels for Idiomify task

  #11 opened 3 years ago by eubinecto
  1

• `d-1-3` : PIE dataset - annotate the idioms with special tokens and add their definitions to `idioms` artifacts

  #5 opened 3 years ago by eubinecto
  1

• `m-1-3`: The same as `m-1-2`, except that it prints out the special tokens before and after the idioms.

  #9 opened 3 years ago by eubinecto
  1

• `t-1-1` : Saving a pre-trained `BartTokenizer` with the special tokens (`<idiom>` , `</idiom>`)

  #7 opened 3 years ago by eubinecto
  0

• Implementing `m-1-2` - testset, metrics, deploy

  #2 opened 3 years ago by eubinecto
  2

• Implementing `m-1-1` - the first baseline

  #1 opened 3 years ago by eubinecto
  7