Word Segmentation on VLSP 2013
98.3F1 ScoreVi Word Segmentation (NlpHUST)
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Vi Word Segmentation (NlpHUST)Source=NlpHUST2026.03 | 98.3 | 4,500 | |
| Our (VnCoreNLP + BPE)Tokenizer=VnCoreNLP + BPE2026.03 | 98.1 | 4,900 | |
| RDRsegmenter2026.03 | 97.9 | 4,700 | |
| VnCoreNLP2026.03 | 97.8 | 5,000 | |
| Underthesea2026.03 | 97.5 | 2,800 | |
| BlingFire2026.03 | 92 | 3,400 | |
| WordPiece (BERT)Model=BERT2026.03 | 91.2 | 3,000 | |
| SentencePiece (Unigram)Algorithm=Unigram2026.03 | 90.8 | 3,100 | |
| BPE (GPT-2)Model=GPT-22026.03 | 89.4 | 3,200 |