Text-to-Video Retrieval on DiDeMo 12 (full-corpus)
26R@1InternVideo2-1B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| InternVideo2-1BParams=1412M, Search Pool Size=9,3842026.01 | 26 | 46.3 | 54.5 | 2,286 | |
| X-PoolParams=153M, Search Pool Size=9,3842026.01 | 22.5 | 42.6 | 51.6 | 3,726 | |
| CLIP4CLIPParams=151M, Search Pool Size=9,3842026.01 | 19 | 36.9 | 48.2 | 2,282 | |
| GRDRParams=63M, Search Pool Size=9,3842026.01 | 15.5 | 29.7 | 36.1 | 119 | |
| T2VIndexerParams=144M, Search Pool Size=9,3842026.01 | 13 | 23.9 | 28.9 | 205 | |
| EERCFParams=164M, Search Pool Size=9,3842026.01 | 10.4 | 27.2 | 35.3 | 4,329 | |
| AVGParams=69M, Search Pool Size=9,3842026.01 | 4.5 | 7.5 | 8.7 | 220 | |
| TIGERParams=69M, Search Pool Size=9,3842026.01 | 3.9 | 6.6 | 7.7 | 243 |