MMRet-MLLM

Reported on 6 benchmarks across 2 tasks · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision6 results

Image RetrievalonCIRCO
mAP@10· 2024-12-19
43.4
SOTA
MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval arXiv:2412.14475
Composed Image Retrieval (CoIR)onCIRCO
mAP@10· 2024-12-19
43.4
SOTA
MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval arXiv:2412.14475
Image RetrievalonFashion IQ
(Recall@10+Recall@50)/2· 2024-12-19
46.1
best: 71.77 (DQU-CIR)
MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval arXiv:2412.14475
Image RetrievalonFashion IQ
Recall@10· 2024-12-19
35.6
best: 61.97 (DQU-CIR)
MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval arXiv:2412.14475
Image RetrievalonCIRR
(Recall@5+Recall_subset@1)/2· 2024-12-19
75.7
best: 83.46 (TMCIR)
MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval arXiv:2412.14475
Image RetrievalonCIRR
Recall@10· 2024-12-19
85.1
best: 91.06 (TMCIR)
MegaPairs: Massive Data Synthesis For Universal Multimodal Retrieval arXiv:2412.14475