Sign language subtitle alignment on WMT-SLT (val)
75.66F1@0.50Segmentation
Evaluation results
| Method | Links | |
|---|---|---|
| SegmentationAlignment Approach=Segment and Align2025.12 | 75.66 | |
| SignCLIP-BSL/ASL/DSGS-finetunedAlignment Approach=Segment, Embed, and Align, Strategy=finetuned2025.12 | 75.34 | |
| SATModel Type=End-to-end trained2025.12 | 75.32 | |
| SignCLIP-multilingualAlignment Approach=Segment, Embed, and Align, Strategy=multilingual2025.12 | 75.16 | |
| Original+ (fixed offsets from training data)Alignment Approach=Fixed offsets2025.12 | 74.17 | |
| Original alignment (usually audio-based)Alignment Approach=Audio-based2025.12 | 47.83 |