Issues

• F1 is not zeroed for an empty candidate/reference when a caller-supplied idf dict gives CLS/SEP nonzero weight

  Tiiiger/bert_score#207 · Filter by author shaurya416 opened on Sep 24, 2026

• OverflowError when using fast tokenizers with models that don't define model_max_length

  Tiiiger/bert_score#205 · Filter by author anaghasimha opened on Jul 29, 2026
  1

• Question: Using modern embedding models as the backbone for BERTScore

  Tiiiger/bert_score#204 · Filter by author Davidwhw opened on May 21, 2026

• Support for more recent models

  Tiiiger/bert_score#202 · Filter by author osmalpkoras opened on Nov 6, 2025

• ERROR: padding_idx must be within num_embeddings

  Tiiiger/bert_score#200 · Filter by author Jacoblincc opened on May 12, 2025
  2

• Discrepancy Between BERTScore CLI and Python API Results

  Tiiiger/bert_score#199 · Filter by author FranHG opened on Apr 15, 2025
  2

• Bert-score gives a high score, even though the two texts are not similar at all

  Tiiiger/bert_score#198 · Filter by author zhu-CX opened on Mar 26, 2025
  1

• Different results for different lenght of input

  Tiiiger/bert_score#197 · Filter by author Ascisel opened on Feb 27, 2025
  1

• EnvironmentError

  Tiiiger/bert_score#190 · Filter by author JingsenZhang opened on Jul 22, 2024
  1

• RuntimeError

  Tiiiger/bert_score#188 · Filter by author fkurushin opened on Jun 14, 2024

• Report: MPS backend is working fine

  Tiiiger/bert_score#187 · Filter by author KanTakahiro opened on Jun 6, 2024
  1

• Scoring Question

  Tiiiger/bert_score#186 · Filter by author schbrough opened on May 13, 2024

• Mabey we cannot assert 'model_type' as a local path?

  Tiiiger/bert_score#185 · Filter by author 141forever opened on May 8, 2024
  6

• Using Clinical Models as Model_Type in BertScore

  Tiiiger/bert_score#184 · Filter by author behnazeslami opened on May 6, 2024

• AttributeError: 'RobertaTokenizerFast' object has no attribute 'max_len'

  Tiiiger/bert_score#183 · Filter by author daehuikim opened on Apr 26, 2024

• is tokenizer max length correct?

  Tiiiger/bert_score#182 · Filter by author ruiguo-bio opened on Apr 26, 2024
  1

• BERTScore can match contextualized embeddings of [SEP]/[CLS] tokens

  Tiiiger/bert_score#180 · Filter by author asumagic opened on Mar 19, 2024

• Question about human ratings

  Tiiiger/bert_score#178 · Filter by author jbdel opened on Mar 15, 2024

• BERT simple true / false comparison scoring seems wrong

  Tiiiger/bert_score#177 · Filter by author james-deee opened on Feb 1, 2024

• DeBERTaV3 Support

  Tiiiger/bert_score#176 · Filter by author TweedBeetle opened on Jan 22, 2024
  1

• Cache embeddings

  Tiiiger/bert_score#175 · Filter by author djaym7 opened on Dec 28, 2023
  2

• Tensor dimension mismatch when using lang='en-sci'

  Tiiiger/bert_score#174 · Filter by author isabelcachola opened on Dec 28, 2023
  2

• Nan of P score?

  Tiiiger/bert_score#173 · Filter by author kevinkexu opened on Oct 28, 2023

• Why is matplotlib needed?

  Tiiiger/bert_score#172 · Filter by author ogencoglu opened on Oct 11, 2023

• Coco challenge link expired

  Tiiiger/bert_score#171 · Filter by author lydia7632 opened on Oct 7, 2023