Text-to-Video Retrieval on MSR-VTT zero-shot
53.9R@1Gemini Embedding 2
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini Embedding 2Setting=ZS2026.05 | 53.9 | |
| InternVideo2-1BSetting=ZS2026.05 | 51.9 | |
| X-CLIPSetting=FT2026.05 | 49.3 | |
| PE-core familySetting=ZS2026.05 | 47.6 | |
| OmniRetriever-7BSetting=ZS2026.05 | 47.6 | |
| LanguageBind (CLIP-H/14)Setting=ZS2026.05 | 44.8 | |
| CLIP4ClipSetting=FT2026.05 | 44.5 | |
| SigLIP-2 familySetting=ZS2026.05 | 38.5 | |
| ImageBind (ViT-H)Setting=ZS2026.05 | 36.1 | |
| Omni-Embed-NemotronSetting=ZS2026.05 | 35.8 |