Video Topic Segmentation on AVLecture (test)
60.01F1 ScoreMMVTS Co-Attn
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| MMVTS Co-AttnPT=true, FT-Coh=true, Modality=V+T2024.08 | 60.01 | 73.88 | 67.27 | 72.32 | 68.37 | — | |
| MMVTS Co-Attn with MoEPT=true, FT-Coh=true, Modality=V+T2024.08 | 59.77 | 75.01 | 67.94 | 71.69 | 68.61 | — | |
| MMVTS Baseline3PT=true, FT-Coh=true, Modality=V+T2024.08 | 58.77 | 72.55 | 67.26 | 71.52 | 67.52 | — | |
| MMVTS Merge-Attn with MoEPT=true, FT-Coh=true, Modality=V+T2024.08 | 57.54 | 73.48 | 64.36 | 70.43 | 66.45 | — | |
| MMVTS Merge-AttnPT=true, FT-Coh=true, Modality=V+T2024.08 | 57.36 | 74.96 | 65.3 | 70.15 | 66.94 | — | |
| MMVTS Baseline2PT=false, FT-Coh=true, Modality=V+T2024.08 | 56.72 | 72.56 | 63.03 | 67.97 | 65.07 | — | |
| MMVTS Baseline1PT=false, FT-Coh=false, Modality=V+T2024.08 | 55.19 | 71.76 | 61.19 | 66.39 | 63.63 | — | |
| LongFormer_csslModality=T2024.08 | 54.02 | 71.56 | 62.4 | 68.39 | 64.09 | — | |
| SWST_seqModality=V+T2024.08 | 53.45 | 70.95 | 59.73 | 65.21 | 62.33 | — | |
| LongFormerModality=T2024.08 | 52.91 | 69.25 | 60.38 | 67.54 | 62.52 | — | |
| Llama-3-8B GenerativeModality=T2024.08 | 40 | 57.55 | 56.52 | 62.8 | 54.22 | — | |
| Llama-3-8B DiscreteModality=T2024.08 | 39.27 | 68.8 | 62.55 | 70.43 | 60.26 | — | |
| MultiSegModel=MultiSeg, Modality=Multi-modal, Language=English2026.02 | 26.46 | — | — | — | — | 17.02 | |
| MiniSeg+Model=MiniSeg+, Modality=Text-only, Language=English2026.02 | 24.64 | — | — | — | — | 15.97 | |
| BaSSLModality=V2024.08 | — | 43.94 | — | 46.95 | — | — | |
| UnsupAVLSModality=V+T2024.08 | — | 56 | — | 70.86 | — | — |