Natural Language Inference on ViNLI
95.78AccuracyHuman
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Human2024.03 | 95.78 | 95.79 | |
| CafeBERTLanguage Type=Monolingual2024.03 | 86.11 | 86.16 | |
| XLM-RobertalargeLanguage Type=Multilingual, Model Variant=Large2024.03 | 85.99 | 86.1 | |
| ViCLSR2026.03 | 82.86 | 82.84 | |
| CafeBERT2026.03 | 82.06 | 82.05 | |
| XLM-Rvariant=Large2026.03 | 81.36 | 81.31 | |
| PhoBERTlargeLanguage Type=Monolingual, Model Variant=Large2024.03 | 80.67 | 80.69 | |
| PhoBERTbaseLanguage Type=Monolingual, Model Variant=Base2024.03 | 78 | 78.05 | |
| XLM-RobertabaseLanguage Type=Multilingual, Model Variant=Base2024.03 | 76.83 | 77.01 | |
| PhoBERTvariant=Large2026.03 | 75.93 | 75.87 | |
| mBERTLanguage Type=Multilingual2024.03 | 73.45 | 73.62 | |
| PhoBERTvariant=Base2026.03 | 72.87 | 72.79 | |
| XLM-Rvariant=Base2026.03 | 71.59 | 71.51 | |
| wikiBERTLanguage Type=Monolingual2024.03 | 71.18 | — | |
| mBERT2026.03 | 64.84 | 64.83 | |
| DiffCSE2026.03 | 58.35 | 58.34 | |
| DistilBERTLanguage Type=Multilingual2024.03 | 44.39 | 66.77 |