T-to-V Retrieval on ANET
70.5Recall@1VAST
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| VASTmodality track=Multi-modal2023.05 | 70.5 | — | — | — | |
| Method [54]modality track=Vision-only2023.05 | 66.8 | — | — | — | |
| Method [7]modality track=Multi-modal2023.05 | 63.4 | — | — | — | |
| UMT (Base)backbone=UMT-L, pretraining_pairs=17M, eval_protocol=zero-shot2024.03 | 41.8 | 267.8 | 68.6 | 79.2 | |
| vid-TLDRbackbone=UMT-L, pretraining_pairs=25M, eval_protocol=zero-shot2024.03 | 41.8 | 152.7 | 68.4 | 79 | |
| vid-TLDRbackbone=UMT-L, pretraining_pairs=17M, eval_protocol=zero-shot2024.03 | 41.7 | 151.7 | 68.8 | 79.1 | |
| UMT + ToMebackbone=UMT-L, pretraining_pairs=25M, eval_protocol=zero-shot2024.03 | 41 | 182.1 | 67.8 | 78.4 | |
| UMT (Base)backbone=UMT-L, pretraining_pairs=25M, eval_protocol=zero-shot2024.03 | 40.7 | 267.8 | 67.9 | 79.3 | |
| UMT + ToMebackbone=UMT-L, pretraining_pairs=17M, eval_protocol=zero-shot2024.03 | 40.3 | 182.1 | 67.1 | 78.2 | |
| vid-TLDRbackbone=UMT-B, pretraining_pairs=25M, eval_protocol=zero-shot2024.03 | 35.7 | 51 | 60.9 | 72 | |
| UMT (Base)backbone=UMT-B, pretraining_pairs=25M, eval_protocol=zero-shot2024.03 | 34.2 | 78.5 | 59.1 | 70.5 | |
| vid-TLDRbackbone=UMT-B, pretraining_pairs=17M, eval_protocol=zero-shot2024.03 | 33.9 | 45 | 59.6 | 70.7 | |
| UMT (Base)backbone=UMT-B, pretraining_pairs=17M, eval_protocol=zero-shot2024.03 | 32.7 | 78.5 | 57.7 | 69.2 | |
| UMT + ToMebackbone=UMT-B, pretraining_pairs=25M, eval_protocol=zero-shot2024.03 | 32.6 | 62.8 | 58.1 | 69.5 | |
| UMT + ToMebackbone=UMT-B, pretraining_pairs=17M, eval_protocol=zero-shot2024.03 | 31.7 | 62.8 | 57.4 | 68.9 | |
| vid-TLDRbackbone=UMT-L, pretraining_pairs=5M, eval_protocol=zero-shot2024.03 | 31.3 | 162 | 59.6 | 71.6 | |
| UMT (Base)backbone=UMT-L, pretraining_pairs=5M, eval_protocol=zero-shot2024.03 | 31 | 267.8 | 59.7 | 71.7 | |
| UMT + ToMebackbone=UMT-L, pretraining_pairs=5M, eval_protocol=zero-shot2024.03 | 29.7 | 182.1 | 57.9 | 70.2 | |
| vid-TLDRbackbone=UMT-B, pretraining_pairs=5M, eval_protocol=zero-shot2024.03 | 28.1 | 49 | 52.4 | 64.1 | |
| UMT (Base)backbone=UMT-B, pretraining_pairs=5M, eval_protocol=zero-shot2024.03 | 27.1 | 78.5 | 51.6 | 63 | |
| UMT + ToMebackbone=UMT-B, pretraining_pairs=5M, eval_protocol=zero-shot2024.03 | 24.9 | 66.2 | 49.1 | 61 |