flairNLP/fabricator

Issues

• Failing CI pytest

  #85 opened a year ago by HallerPatrick
  0

• Supporting other open-source LLMs

  #83 opened a year ago by mzeidhassan
  2

• Remove spacy dependency

  #55 opened a year ago by whoisjones
  0

• Initial version for pypi package

  #73 opened a year ago by whoisjones
  0

• Naming of sampling strategies

  #77 opened a year ago by whoisjones
  1

• after accepted: create pypi package

  #6 opened a year ago by fhamborg
  1

• Cannot import DatasetGenerator

  #74 opened a year ago by demongolem-biz2
  1

• Large Git Pack Files

  #71 opened a year ago by HallerPatrick
  2

• Implement ProGen approach

  #67 opened a year ago by whoisjones
  2

• Add Fabriator Predictions to Wrench Datasets

  #70 opened a year ago by alanakbik
  0

• Guided Generation for Syntax dependent tasks

  #69 opened a year ago by HallerPatrick
  0

• Idea on how to structure generation / annotation

  #68 opened a year ago by whoisjones
  1

• Add predicted labels to CoNLL-03 dataset

  #66 opened a year ago by alanakbik
  1

• Improve fewshot sampling naming convention

  #58 opened a year ago by whoisjones
  0

• Rename to Fabricator

  #60 opened a year ago by whoisjones
  0

• improve logo + fix width in readme

  #62 opened a year ago by whoisjones
  0

• Split haystack dependencies

  #56 opened a year ago by whoisjones
  2

• for evaluation: compare against baseline that just k times repeats the few shot examples

  #22 opened a year ago by fhamborg
  0

• severe: class_encode_column does not convert to same original int labels

  #51 opened a year ago by fhamborg
  0

• Reduce dependencies

  #52 opened a year ago by whoisjones
  1

• enable user to provide 1-sentence description for each label, which will be shown at the beginning of the prompt to the LLM

  #17 opened a year ago by fhamborg
  2

• Refactorings for submission

  #42 opened a year ago by whoisjones
  0

• Custom Prefixes for data column and few shot column for prompt

  #45 opened a year ago by HallerPatrick
  2

• Rename source directory to "src"

  #3 opened a year ago by HallerPatrick
  0

• More log infos

  #35 opened a year ago by HallerPatrick
  0

• cost reduction: generate and annotate within a single prompt (instead of first generating and then annotating)

  #29 opened a year ago by fhamborg
  0

• execution speed degrades with increasing sizes of datasets to be annotated

  #48 opened a year ago by fhamborg
  3

• Ensure robust and fault tolerant generation

  #9 opened a year ago by HallerPatrick
  0

• Make it possible to use all prompts with text inputs rather than label IDs

  #14 opened a year ago by whoisjones
  0

• How to properly differentiate between generate unlabeled and annotate unlabeled data?

  #11 opened a year ago by whoisjones
  2

• Naming convention for DataPoints / Prompts

  #8 opened a year ago by whoisjones
  0

• Sampling method that covers all labels

  #5 opened 2 years ago by HallerPatrick
  0