Sign Language Translation on CSL-Daily
51.26ROUGE-LLLAVA-SLT
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LLAVA-SLTVLP (Vision-Language Pre-training)=true2026.04 | 51.26 | 52.15 | 36.24 | 26.47 | 20.42 | — | |
| SCL-SLTVLP (Vision-Language Pre-training)=true2026.04 | 51.08 | 52.81 | 39.28 | 29.82 | 23.25 | — | |
| GFSLT-VLP-SignCLVLP (Vision-Language Pre-training)=true2026.04 | 48.92 | 47.47 | 32.53 | 22.62 | 16.16 | — | |
| MMSLTVLP (Vision-Language Pre-training)=true2026.04 | 48.92 | 49.87 | 36.37 | 27.29 | 21.11 | — | |
| SCL-SLTVLP (Vision-Language Pre-training)=false2026.04 | 48.53 | 50.36 | 36.88 | 27.69 | 21.41 | — | |
| C2RLVLP (Vision-Language Pre-training)=true2026.04 | 48.21 | 49.32 | 36.28 | 27.54 | 21.61 | — | |
| SpaMoVLP (Vision-Language Pre-training)=true2026.04 | 47.46 | 48.9 | 36.9 | 26.78 | 20.55 | — | |
| Sign2GPTVLP (Vision-Language Pre-training)=false2026.04 | 42.36 | 41.75 | 28.73 | 20.6 | 15.4 | — | |
| SignLLMVLP (Vision-Language Pre-training)=false2026.04 | 39.91 | 39.55 | 28.13 | 20.07 | 15.75 | — | |
| FLa-LLMVLP (Vision-Language Pre-training)=false2026.04 | 37.25 | 37.13 | 25.12 | 18.38 | 14.2 | — | |
| GFSLT-VLPVLP (Vision-Language Pre-training)=true2026.04 | 36.44 | 39.37 | 24.93 | 16.26 | 11 | — | |
| GFSLTVLP (Vision-Language Pre-training)=false2026.04 | 35.16 | 37.69 | 23.28 | 14.93 | 9.88 | — | |
| NSLT+LuongVLP (Vision-Language Pre-training)=false2026.04 | 34.54 | 34.16 | 19.57 | 11.84 | 7.56 | — | |
| MMTLBVLP (Vision-Language Pre-training)=false2026.04 | 26.7 | 27.22 | 15.9 | 10.61 | 7.61 | — | |
| GASLTgloss-free=true2023.07 | 20.35 | 19.9 | 9.94 | 5.98 | 4.07 | — | |
| GASLTVLP (Vision-Language Pre-training)=false2026.04 | 20.35 | 19.9 | 9.94 | 5.98 | 4.07 | — | |
| SLRTVLP (Vision-Language Pre-training)=false2026.04 | 19.67 | 20 | 9.11 | 4.93 | 3.03 | — | |
| Joint-SLTgloss-free=true2023.07 | 19.61 | 21.56 | 8.29 | 3.68 | 1.72 | — | |
| TSPNet-Jointgloss-free=true2023.07 | 18.38 | 17.09 | 8.98 | 5.07 | 2.97 | — | |
| Self-Mutual KDFeature Type=Gloss-based, Translator=Joint-SLT2024.05 | — | — | — | — | — | 11.61 | |
| Self-Mutual KDFeature Type=Gloss-based, Translator=NSLT2024.05 | — | — | — | — | — | 8.97 | |
| Self-Mutual KD + SignCLFeature Type=Gloss-based, Translator=NSLT, SignCL Training Strategy=applied to NSLT2024.05 | — | — | — | — | — | 10.35 | |
| VLP PretrainedFeature Type=Gloss-free, Translator=Joint-SLT2024.05 | — | — | — | — | — | 2.93 | |
| VLP PretrainedFeature Type=Gloss-free, Translator=NSLT2024.05 | — | — | — | — | — | 1.82 | |
| VLP Pretrained + SignCLFeature Type=Gloss-free, Translator=NSLT, SignCL Training Strategy=applied to NSLT2024.05 | — | — | — | — | — | 2.19 | |
| VLP Pretrained + SignCLFeature Type=Gloss-free with SignCL extraction, Translator=Joint-SLT2024.05 | — | — | — | — | — | 3.18 | |
| VLP Pretrained + SignCLFeature Type=Gloss-free with SignCL extraction, Translator=NSLT2024.05 | — | — | — | — | — | 2.29 | |
| VLP Pretrained + SignCL + SignCLFeature Type=Gloss-free with SignCL extraction, Translator=NSLT, SignCL Training Strategy=applied to NSLT2024.05 | — | — | — | — | — | 2.53 |