Moment Retrieval on QVHighlights (val)
78.65R@1 (IoU=0.5)SMART
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| SMARTBackbone=CLIP+LLM2025.11 | 78.65 | 65.03 | 70.46 | 56.72 | — | — | 72.03 | |
| LLaVA-MRBackbone=BLIP-22024.11 | 78.13 | 64.13 | 69.64 | 56.32 | — | — | 71.85 | |
| LLaVA-MRBackbone=CLIP+LLM2025.11 | 78.13 | 64.13 | 69.64 | 56.32 | — | — | 71.85 | |
| Mr. BLIPBackbone=BLIP-22024.11 | 76.13 | 63.35 | 69.39 | 55.78 | — | — | — | |
| Mr.BLIPBackbone=BLIP-22025.11 | 76.13 | 63.35 | 69.39 | 55.78 | — | — | — | |
| FlashVTGBackbone=SF+CLIP2025.11 | 73.1 | 57.29 | 72.75 | 54.33 | — | — | — | |
| VideoLights-B-ptBackbone=SF+CLIP+BLIP2025.11 | 72.06 | 57.94 | 70.38 | 51.12 | — | — | — | |
| InternVideo2-6BFeature=InternVideo2_S2-6B2024.03 | 71.42 | 56.45 | — | — | — | 49.24 | — | |
| InternVideo2-1BFeature=InternVideo2_S2-1B2024.03 | 70 | 54.45 | — | — | — | 47.02 | — | |
| R2-TuningBackbone=CLIP2025.11 | 68.71 | 52.06 | — | — | — | — | — | |
| CG-DETRFeatures=Slowfast + CLIP2023.11 | 67.4 | 52.1 | 65.6 | 45.7 | 44.9 | — | — | |
| CG-DETRBackbone=SF+CLIP2024.11 | 67.4 | 52.1 | 65.6 | 45.7 | — | — | — | |
| CG-DETRBackbone=SF+CLIP2025.11 | 67.4 | 52.1 | 65.6 | 45.7 | — | — | — | |
| CG-DETR2024.04 | 67.35 | 52.06 | 65.57 | 45.73 | 44.93 | — | — | |
| UnLoc-LVision Enc.=CLIP2023.08 | 66.1 | 46.7 | — | — | — | — | — | |
| UnLoc-LBackbone=CLIP*2024.11 | 66.1 | 46.7 | — | — | — | — | — | |
| UnLoc-LBackbone=CLIP*2025.11 | 66.1 | 46.7 | — | — | — | — | — | |
| CLIP+SlowFastFeature=CLIP+SlowFast2024.03 | 65.43 | 48.38 | — | — | — | 42.86 | — | |
| BAM-DETRFLOPs=0.65G, Params=9.5M2023.11 | 65.1 | 51.81 | 65.41 | 48.56 | 47.61 | — | — | |
| UVCOMBackbone=SF+CLIP2025.11 | 65.1 | 51.81 | — | — | — | — | — | |
| CLIPFeature=CLIP2024.03 | 64.97 | 48.65 | — | — | — | 42.96 | — | |
| UnLoc-BVision Enc.=CLIP2023.08 | 64.5 | 48.8 | — | — | — | — | — | |
| UnLocBackbone=CLIP-B/16, Post Pre-train=650K Videos, #Params=87.9M*2024.03 | 64.5 | 48.8 | — | — | — | — | — | |
| TaskWeave2024.04 | 64.26 | 50.06 | 65.39 | 46.47 | 45.38 | — | — | |
| BAM-DETRslimFLOPs=0.43G, Params=7.2M2023.11 | 63.94 | 50.19 | 64.51 | 48.51 | 47.03 | — | — | |
| GroundVTS-IBackbone=InternVL3.5-8B2026.04 | 63.6 | 40.7 | — | — | — | — | — | |
| QD-DETRFLOPs=0.59G, Params=7.7M2023.11 | 62.9 | 46.77 | 62.66 | 41.51 | 41.24 | — | — | |
| QD-DETRFeatures=Slowfast + CLIP2023.11 | 62.7 | 46.7 | 62.2 | 41.8 | 41.2 | — | — | |
| QD-DETR2024.04 | 62.68 | 46.66 | 62.23 | 41.82 | 41.22 | — | — | |
| QD-DETR2024.04 | 62.68 | 46.66 | 62.23 | 41.82 | 41.22 | — | — | |
| QD-DETRBackbone=SF+CLIP2024.11 | 62.68 | 46.66 | 62.23 | 41.82 | — | — | — | |
| QD-DETRBackbone=SF+CLIP2025.11 | 62.68 | 46.66 | 62.23 | 41.82 | — | — | — | |
| QD-DETRVision Enc.=SF+CLIP2023.08 | 62.4 | 45 | — | — | — | — | — | |
| GranAlignMLLM=true, Setting=Zero-shot2026.01 | 61.94 | 41.81 | 59.63 | — | 39.12 | — | — | |
| EaTRFeatures=Slowfast + CLIP2023.11 | 61.4 | 45.8 | 61.9 | 41.9 | 41.7 | — | — | |
| EaTR2024.04 | 61.36 | 45.79 | 61.86 | 41.91 | 41.74 | — | — | |
| EaTR2024.04 | 61.36 | 45.79 | 61.86 | 41.91 | 41.74 | — | — | |
| EaTRBackbone=SlowFast R-50, CLIP-B/32, PANN, Post Pre-train=No, #Params=9.0M2024.03 | 61.36 | 45.79 | 61.86 | 41.91 | 41.74 | — | — | |
| EaTRBackbone=I3D2024.11 | 61.36 | 45.79 | 61.86 | 41.91 | — | — | — | |
| EaTRBackbone=I3D2025.11 | 61.36 | 45.79 | 61.86 | 41.91 | — | — | — | |
| EaTRFLOPs=0.47G, Params=9.1M2023.11 | 60.9 | 46.13 | 62.01 | 42.17 | 41.43 | — | — | |
| UMTVision Enc.=SF+CLIP2023.08 | 60.3 | 44.3 | — | — | — | — | — | |
| UMTFeatures=Slowfast + CLIP2023.11 | 60.3 | 44.3 | — | — | 38.6 | — | — | |
| UMT2024.04 | 60.26 | 44.26 | — | — | 38.59 | — | — | |
| UMTaudio modality=true2024.04 | 60.26 | 44.26 | — | — | 38.59 | — | — | |
| UniVGT2024.04 | 59.74 | — | — | — | 36.13 | — | — | |
| UniVTGpretrain-finetune paradigm=true2024.04 | 59.74 | — | — | — | 36.13 | — | — | |
| UniVTG++FLOPs=0.98G, Params=43.4M2023.11 | 59.74 | 40.9 | 58.61 | 36.76 | 36.13 | — | — | |
| UniVTGFeatures=Slowfast + CLIP2023.11 | 59.7 | 41.8 | — | — | 36.1 | — | — | |
| Moment-DETRBackbone=SF+CLIP2024.11 | 59.68 | 40.84 | — | — | — | — | — | |
| Moment-DETRBackbone=SF+CLIP2025.11 | 59.68 | 40.84 | — | — | — | — | — | |
| Moment-GPTMLLM=true, Setting=Zero-shot2026.01 | 58.9 | 38.6 | 55.7 | — | 35.9 | — | — | |
| UniVTG2026.04 | 58.9 | 40.9 | — | — | — | — | — | |
| MomentDiffMLLM=false, Setting=Fully-supervised2026.01 | 57.8 | 39.2 | 54.6 | — | 35.3 | — | — | |
| SeViLA2026.04 | 54.5 | 36.5 | — | — | — | — | — | |
| Moment-DETRMLLM=false, Setting=Fully-supervised2026.01 | 54.2 | 33.4 | 55.4 | — | 31.1 | — | — | |
| Moment-DETR2024.04 | 53.94 | 34.84 | — | — | 32.2 | — | — | |
| Moment-DETR2024.04 | 53.94 | 34.84 | — | — | 32.2 | — | — | |
| M-DETRVision Enc.=SF+CLIP2023.08 | 53.9 | 34.8 | — | — | — | — | — | |
| M-DETRFeatures=Slowfast + CLIP2023.11 | 53.9 | 34.8 | — | — | 32.2 | — | — | |
| LLaViLoMLLM=true, Setting=Fully-supervised2026.01 | 49 | 30.4 | 49.4 | — | 28.9 | — | — | |
| VTimeLLMMLLM=true, Setting=Fully-supervised2026.01 | 48.8 | 29.5 | 49.3 | — | 26.8 | — | — | |
| DiwanMLLM=true, Setting=Zero-shot2026.01 | 48.3 | 31 | 47.3 | — | 28 | — | — | |
| VCL2024.04 | 43.33 | 25.75 | 39.23 | 24.95 | 21.13 | — | — | |
| InternVL3.5-8B-GBackbone=InternVL3.5-8B, SFT=Grounding-FT2026.04 | 31.8 | 15 | — | — | — | — | — | |
| GroundVTS-QBackbone=Qwen2.5VL-7B2026.04 | 23.6 | 12.3 | — | — | — | — | — | |
| VideoChatGPTMLLM=true, Setting=Zero-shot2026.01 | 22.4 | — | 21.9 | — | 10.3 | — | — | |
| VideoLLaMaMLLM=true, Setting=Zero-shot2026.01 | 18.5 | 6.9 | 17.8 | — | 7.1 | — | — | |
| PZVMRMLLM=false, Setting=Unsupervised2026.01 | 12.6 | 5.1 | 16.2 | — | 5.3 | — | — | |
| Liu et al.MLLM=false, Setting=Unsupervised2026.01 | 12.3 | 3.5 | 10.4 | — | 2.7 | — | — | |
| Qwen2.5VL-7B-GBackbone=Qwen2.5VL-7B, SFT=Grounding-FT2026.04 | 11 | 4.3 | — | — | — | — | — | |
| Qwen2.5VL-7BBackbone=Qwen2.5VL-7B2026.04 | 8.7 | 2.4 | — | — | — | — | — | |
| InternVL3.5-8BBackbone=InternVL3.5-8B2026.04 | 8.7 | 3.7 | — | — | — | — | — |