Video-to-Text Retrieval on ActivityNet (val1)
46.7R@1CenterCLIP
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CenterCLIPBackbone=ViT-B/16, Token Clustering Strategy=k-medoids++, Clustering Settings (Ba - S, K)=15, 1602022.05 | 46.7 | 77.1 | 88 | 2 | 5.5 | — | |
| CenterCLIPBackbone=ViT-B/32, Token Clustering Strategy=k-medoids++, Clustering Settings (Ba - S, K)=12, 492022.05 | 44.5 | 75.3 | 86 | 2 | 6.7 | — | |
| CenterCLIPBackbone=ViT-B/32, Token Clustering Strategy=spectral, Clustering Settings (Ba - S, K)=15, 492022.05 | 44.5 | 75.7 | 86.2 | 2 | 6.5 | — | |
| X-CLIP2024.01 | 44.4 | 74.7 | 86.3 | — | — | 68.5 | |
| CenterCLIPBackbone=ViT-B/32, Token Clustering Strategy=k-medoids++, Clustering Settings (Ba - S, K)=15, 492022.05 | 44.2 | 75 | 86.1 | 2 | 6.8 | — | |
| CenterCLIPBackbone=ViT-B/32, Token Clustering Strategy=spectral, Clustering Settings (Ba - S, K)=20, 492022.05 | 44.1 | 75.1 | 86 | 2 | 6.7 | — | |
| CenterCLIP2024.01 | 44.1 | 75.1 | 86 | — | — | 68.4 | |
| EERCF2024.01 | 42.9 | 76.2 | 87.7 | — | — | 68.9 | |
| CLIP4clip (meanP) baselineBackbone=ViT-B/322022.05 | 42.8 | 73.8 | 85.3 | 2 | 6.9 | — | |
| CLIP4clip (meanP)2022.05 | 42.5 | 74.1 | 85.8 | 2 | 6.6 | — | |
| DRL2024.01 | 42.2 | 74 | 86.2 | — | — | 67.5 | |
| CLIP4clip (seqTransf)2022.05 | 41.4 | 73.7 | 85.3 | 2 | 6.7 | — | |
| TS2Net2024.01 | 41.3 | 72.1 | 85.1 | — | — | 66.2 | |
| CLIP4Clip2024.01 | 39.5 | 71.4 | 83.6 | — | — | 64.8 | |
| MMT2022.05 | 28.9 | 61.1 | — | 4 | 17.1 | — | |
| SSB2022.05 | 28.7 | 60.8 | — | 2 | — | — | |
| CMGSD2022.05 | 24.6 | 56.8 | — | 4 | — | — | |
| T2VLAD2022.05 | 24.1 | 56.6 | — | 4 | — | — | |
| TT-CE+2022.05 | 23 | 56.1 | — | 4 | — | — | |
| TeachText2024.01 | 23 | 56.1 | 95.8 | — | — | 58.3 | |
| CLIPzero-shot=true2022.05 | 17.9 | 40.8 | 54.2 | 8 | 43.3 | — | |
| CE2022.05 | 17.7 | 46.6 | — | 6 | 24.4 | — | |
| FSE2022.05 | 16.7 | 43.1 | — | 7 | — | — |