List your company

Moment Retrieval on QVHighlights (val)

78.65R@1 (IoU=0.5)SMART
5.90224.788543.67562.5615Aug 21, 2023Jan 27, 2024Jul 4, 2024Dec 10, 2024May 18, 2025Oct 24, 2025Apr 2, 2026

Evaluation results

MethodLinks
SMARTBackbone=CLIP+LLM2025.1178.6565.0370.4656.72——72.03
LLaVA-MRBackbone=BLIP-22024.1178.1364.1369.6456.32——71.85
LLaVA-MRBackbone=CLIP+LLM2025.1178.1364.1369.6456.32——71.85
Mr. BLIPBackbone=BLIP-22024.1176.1363.3569.3955.78———
Mr.BLIPBackbone=BLIP-22025.1176.1363.3569.3955.78———
FlashVTGBackbone=SF+CLIP2025.1173.157.2972.7554.33———
VideoLights-B-ptBackbone=SF+CLIP+BLIP2025.1172.0657.9470.3851.12———
InternVideo2-6BFeature=InternVideo2_S2-6B2024.0371.4256.45———49.24—
InternVideo2-1BFeature=InternVideo2_S2-1B2024.037054.45———47.02—
R2-TuningBackbone=CLIP2025.1168.7152.06—————
CG-DETRFeatures=Slowfast + CLIP2023.1167.452.165.645.744.9——
CG-DETRBackbone=SF+CLIP2024.1167.452.165.645.7———
CG-DETRBackbone=SF+CLIP2025.1167.452.165.645.7———
CG-DETR2024.0467.3552.0665.5745.7344.93——
UnLoc-LVision Enc.=CLIP2023.0866.146.7—————
UnLoc-LBackbone=CLIP*2024.1166.146.7—————
UnLoc-LBackbone=CLIP*2025.1166.146.7—————
CLIP+SlowFastFeature=CLIP+SlowFast2024.0365.4348.38———42.86—
BAM-DETRFLOPs=0.65G, Params=9.5M2023.1165.151.8165.4148.5647.61——
UVCOMBackbone=SF+CLIP2025.1165.151.81—————
CLIPFeature=CLIP2024.0364.9748.65———42.96—
UnLoc-BVision Enc.=CLIP2023.0864.548.8—————
UnLocBackbone=CLIP-B/16, Post Pre-train=650K Videos, #Params=87.9M*2024.0364.548.8—————
TaskWeave2024.0464.2650.0665.3946.4745.38——
BAM-DETRslimFLOPs=0.43G, Params=7.2M2023.1163.9450.1964.5148.5147.03——
GroundVTS-IBackbone=InternVL3.5-8B2026.0463.640.7—————
QD-DETRFLOPs=0.59G, Params=7.7M2023.1162.946.7762.6641.5141.24——
QD-DETRFeatures=Slowfast + CLIP2023.1162.746.762.241.841.2——
QD-DETR2024.0462.6846.6662.2341.8241.22——
QD-DETR2024.0462.6846.6662.2341.8241.22——
QD-DETRBackbone=SF+CLIP2024.1162.6846.6662.2341.82———
QD-DETRBackbone=SF+CLIP2025.1162.6846.6662.2341.82———
QD-DETRVision Enc.=SF+CLIP2023.0862.445—————
GranAlignMLLM=true, Setting=Zero-shot2026.0161.9441.8159.63—39.12——
EaTRFeatures=Slowfast + CLIP2023.1161.445.861.941.941.7——
EaTR2024.0461.3645.7961.8641.9141.74——
EaTR2024.0461.3645.7961.8641.9141.74——
EaTRBackbone=SlowFast R-50, CLIP-B/32, PANN, Post Pre-train=No, #Params=9.0M2024.0361.3645.7961.8641.9141.74——
EaTRBackbone=I3D2024.1161.3645.7961.8641.91———
EaTRBackbone=I3D2025.1161.3645.7961.8641.91———
EaTRFLOPs=0.47G, Params=9.1M2023.1160.946.1362.0142.1741.43——
UMTVision Enc.=SF+CLIP2023.0860.344.3—————
UMTFeatures=Slowfast + CLIP2023.1160.344.3——38.6——
UMT2024.0460.2644.26——38.59——
UMTaudio modality=true2024.0460.2644.26——38.59——
UniVGT2024.0459.74———36.13——
UniVTGpretrain-finetune paradigm=true2024.0459.74———36.13——
UniVTG++FLOPs=0.98G, Params=43.4M2023.1159.7440.958.6136.7636.13——
UniVTGFeatures=Slowfast + CLIP2023.1159.741.8——36.1——
Moment-DETRBackbone=SF+CLIP2024.1159.6840.84—————
Moment-DETRBackbone=SF+CLIP2025.1159.6840.84—————
Moment-GPTMLLM=true, Setting=Zero-shot2026.0158.938.655.7—35.9——
UniVTG2026.0458.940.9—————
MomentDiffMLLM=false, Setting=Fully-supervised2026.0157.839.254.6—35.3——
SeViLA2026.0454.536.5—————
Moment-DETRMLLM=false, Setting=Fully-supervised2026.0154.233.455.4—31.1——
Moment-DETR2024.0453.9434.84——32.2——
Moment-DETR2024.0453.9434.84——32.2——
M-DETRVision Enc.=SF+CLIP2023.0853.934.8—————
M-DETRFeatures=Slowfast + CLIP2023.1153.934.8——32.2——
LLaViLoMLLM=true, Setting=Fully-supervised2026.014930.449.4—28.9——
VTimeLLMMLLM=true, Setting=Fully-supervised2026.0148.829.549.3—26.8——
DiwanMLLM=true, Setting=Zero-shot2026.0148.33147.3—28——
VCL2024.0443.3325.7539.2324.9521.13——
InternVL3.5-8B-GBackbone=InternVL3.5-8B, SFT=Grounding-FT2026.0431.815—————
GroundVTS-QBackbone=Qwen2.5VL-7B2026.0423.612.3—————
VideoChatGPTMLLM=true, Setting=Zero-shot2026.0122.4—21.9—10.3——
VideoLLaMaMLLM=true, Setting=Zero-shot2026.0118.56.917.8—7.1——
PZVMRMLLM=false, Setting=Unsupervised2026.0112.65.116.2—5.3——
Liu et al.MLLM=false, Setting=Unsupervised2026.0112.33.510.4—2.7——
Qwen2.5VL-7B-GBackbone=Qwen2.5VL-7B, SFT=Grounding-FT2026.04114.3—————
Qwen2.5VL-7BBackbone=Qwen2.5VL-7B2026.048.72.4—————
InternVL3.5-8BBackbone=InternVL3.5-8B2026.048.73.7—————