List your company
ResearchFormerly WizWand

Composed Image Retrieval on CIRCO 1.0 (test)

42.2mAP@5MMRet-MLLM
3.09613.24823.433.552Dec 19, 2024Jan 4, 2025Jan 20, 2025Feb 5, 2025Feb 21, 2025Mar 9, 2025Mar 25, 2025
Updated 8mo ago

Evaluation results

MethodLinks
MMRet-MLLMBackbone=LLaVA-1.6, # Params=7.57B, Zero-shot evaluation=true2024.1242.2——
MMRet-LargeBackbone=CLIP-L, # Params=428M, Zero-shot evaluation=true2024.1239.2——
MMRet-BaseBackbone=CLIP-B, # Params=149M, Zero-shot evaluation=true2024.1234.3——
MagicLens-L*Backbone=CoCa-L, # Params=613M, Zero-shot evaluation=true2024.1234.1——
IP-CIRBackbone=CLIP-G, # Params=43.8B+, Zero-shot evaluation=true2024.1232.8——
MM-EmbedBackbone=LLaVA-1.6, # Params=7.57B, Zero-shot evaluation=true2024.1232.3——
LDREBackbone=CLIP-G, # Params=10.3B+, Zero-shot evaluation=true2024.1231.1——
MagicLens-B*Backbone=CoCa-B, # Params=267M, Zero-shot evaluation=true2024.1230.8——
MagicLens-LBackbone=CLIP-L, # Params=465M, Zero-shot evaluation=true2024.1229.6——
CIREVLBackbone=CLIP-G, # Params=14.6B+, Zero-shot evaluation=true2024.1226.8——
LDREBackbone=CLIP-L, # Params=8.2B+, Zero-shot evaluation=true2024.1223.4——
MagicLens-BBackbone=CLIP-B, # Params=166M, Zero-shot evaluation=true2024.1223.1——
CoLLMVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot, Incorporates LLM=true2025.0320.320.823.4
CoLLMVision Encoder=BLIP-L/16, Evaluation Protocol=Zero-shot, Incorporates LLM=true2025.0319.720.423.1
E5-VBackbone=LLaVA-1.6, # Params=8.35B, Zero-shot evaluation=true2024.1219.1——
CIREVLBackbone=CLIP-L, # Params=12.5B+, Zero-shot evaluation=true2024.1218.6——
CIREVLVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot, Incorporates LLM=true2025.0318.61921.8
Slerp-TATVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot2025.0318.519.421.4
LDREBackbone=CLIP-B, # Params=7.9B+, Zero-shot evaluation=true2024.1218——
MCLVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot, Incorporates LLM=true2025.0317.818.421.8
CIREVLBackbone=CLIP-B, # Params=12.3B+, Zero-shot evaluation=true2024.1214.9——
CIREVLVision Encoder=OpenAI CLIP-B/32, Evaluation Protocol=Zero-shot, Incorporates LLM=true2025.0314.915.417.8
Slerp-TATVision Encoder=BLIP-L/16, Evaluation Protocol=Zero-shot2025.0313.818.421.1
LinCIRVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot, Reproduced Results=true2025.031313.916.2
ContextI2WVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot2025.031313.816
CoLLMVision Encoder=OpenAI CLIP-B/32, Evaluation Protocol=Zero-shot, Incorporates LLM=true2025.0312.913.215
CompoDiffBackbone=CLIP-L, # Params=568M, Zero-shot evaluation=true2024.1212.6——
SEARLEBackbone=CLIP-L, # Params=442M, Zero-shot evaluation=true2024.1211.7——
SEARLEVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot2025.0311.712.715.1
PLIBackbone=CLIP-L, # Params=428M, Zero-shot evaluation=true2024.1210.4——
SEARLEBackbone=CLIP-B, # Params=165M, Zero-shot evaluation=true2024.129.4——
SEARLEVision Encoder=OpenAI CLIP-B/32, Evaluation Protocol=Zero-shot2025.039.49.911.8
Slerp-TATVision Encoder=OpenAI CLIP-B/32, Evaluation Protocol=Zero-shot2025.039.310.312.3
Pic2WordBackbone=CLIP-L, # Params=429M, Zero-shot evaluation=true2024.128.7——
Pic2WorldVision Encoder=OpenAI CLIP-L/14, Evaluation Protocol=Zero-shot2025.038.79.511.3
PALAVRAVision Encoder=OpenAI CLIP-B/32, Evaluation Protocol=Zero-shot2025.034.65.36.8