List your company
ResearchFormerly WizWand

Visual Question Answering on InfoSeek (val)

47.2Overall AccuracyReAG
3.10414.5522637.448Apr 5, 2025May 14, 2025Jun 22, 2025Aug 1, 2025Sep 9, 2025Oct 18, 2025Nov 27, 2025
Updated 5mo ago

Evaluation results

MethodLinks
ReAGGenerator=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B2025.1147.248.346.2
ReflectiVAGenerator=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B2025.1143.943.544.3
OMGMRetriever=EVA-CLIP-8B, Model=LLaVA-1.5-7B, Reranking=true2025.0843.5——
OMGMGenerator Model=LLaVA-1.5-7B, Gen. FT=false, Ret. FT=true2025.0543.4943.4643.53
ReAGGenerator=Qwen2.5-VL-3B, Retriever=EVA-CLIP-8B2025.1143.343.742.9
VLM-PRFGenerator=Qwen2.5-VL-7B, Retriever=EVA-CLIP-8B2025.1142.843.342.7
VLM-PRFGenerator=InternVL3-8B, Retriever=EVA-CLIP-8B2025.1142.543.542.1
Wiki-PRF(w/ RL)Retriever=EVA-CLIP-8B, Model=VLM-PRF-7B2025.0842.5——
OMGMRetriever=EVA-CLIP-8B, Model=GPT-4o, Reranking=true2025.0842.1——
VLM-PRFGenerator=LLaVA-MORE-8B, Retriever=EVA-CLIP-8B2025.1140.841.340.6
mKG-RAGGenerator=LLaVA-MORE-8B, Retriever=Custom VLM2025.1140.541.439.6
mR2AGGenerator Model=LLaVA-1.5-7B, Gen. FT=true, Ret. FT=true2025.0540.240.639.8
mR2AGGenerator=LLaVA-v1.5-7B, Retriever=CLIP ViT-L/142025.1140.240.639.8
mR²AGRetriever=CLIP-ViT-L, Model=LLaVA-1.5-7B, Reranking=true2025.0840.2——
ReflectiVAGenerator Model=LLAVA-MORE-8B, Gen. FT=true, Ret. FT=true2025.0540.140.439.8
ReflectiVAGenerator=LLaVA-MORE-8B, Retriever=EVA-CLIP-8B2025.1140.140.439.8
ReflectiVARetriever=EVA-CLIP-8B, Model=LLaVA-MORE-8B2025.0840.1——
VLM-PRFGenerator=Qwen2.5-VL-3B, Retriever=EVA-CLIP-8B2025.113939.738.8
ReflectiVAGenerator=Qwen2.5-VL-3B, Retriever=EVA-CLIP-8B2025.1138.939.638.1
MMKB-RAGRetriever=PreFLMR ViT-G, Model=Qwen2-VL-7B, Reranking=true2025.0836.7——
OMGMGenerator Model=InternVL-2.5-8B, Gen. FT=false, Ret. FT=true2025.0536.137.1635.1
OMGMGenerator Model=LLaMA3-8B, Gen. FT=false, Ret. FT=true2025.0534.4235.2633.61
EchoSightGenerator Model=Mistral-7B | LLaMA3-8B, Gen. FT=false, Ret. FT=true2025.0531.3——
EchoSightGenerator=Mistral-7B/LLaMA-3-8B, Retriever=EVA-CLIP-8B2025.1131.3——
EchoSightRetriever=EVA-CLIP-8B, Model=Mistral-7B, Reranking=true2025.0831.3——
EchoSightGenerator=LLaMA-3.1-8B, Retriever=EVA-CLIP-8B2025.1130.43030.7
Wiki-LLaVAGenerator Model=LLaVA-1.5-7B, Gen. FT=false, Ret. FT=true2025.0528.930.127.8
Wiki-LLaVAGenerator=LLaVA-v1.5-7B, Retriever=CLIP ViT-L/14+Contriever2025.1128.930.127.8
Wiki-LLaVARetriever=CLIP-ViT-L, Model=LLaVA-1.5-7B2025.0828.9——
Qwen2.5-VL-7B2025.1123.722.824.1
LLM-RAGenerator Model=BLIP2-Flan-T5XL, Gen. FT=true, Ret. FT=true2025.0523.1426.1220.9
LLM-RARetriever=EVA-CLIP-8B, Model=BLIP2-Flan-T5XL2025.0823.1——
RET-2Retriever=RET-2, Model=LLaVA-MORE-8B2025.0822.8——
UniRVQABase Models=InstructBLIP, Evaluation Protocol=Fine-tuned with knowledge, K=52025.0422.0624.0120.4
CLIP + FiDEvaluation Protocol=Fine-tuned with knowledge2025.0420.923.319.1
CLIP + PaLMBase Models=PaLM(540B), Evaluation Protocol=Fine-tuned with knowledge2025.0420.422.718.5
PALI-17BBase Models=PALI, Evaluation Protocol=Standard fine-tuned on the dataset2025.0419.724.216.7
Qwen2.5-VL-3B2025.1118.318.917.7
InstructBLIP-T5XLBase Models=InstructBLIP, Evaluation Protocol=Standard fine-tuned on the dataset2025.0414.51514
BLIP-22025.1112.512.712.3
BLIP2-T5XLBase Models=BLIP2, Evaluation Protocol=Standard fine-tuned on the dataset2025.0412.512.712.3
DPRV+TGenerator=Multi-passage BERT, Retriever=CLIP ViT-B/322025.1112.4——
LLaVA-v1.5-7B2025.119.59.69.4
LLaVA-MORE-8B2025.117.88.38.9
PaLM(Q-only)Base Models=PaLM, Evaluation Protocol=Standard fine-tuned on the dataset2025.044.85.54.2
CoMEMGenerator=Qwen2.5-VL-7B, Retriever=Custom VLM2025.11—32.828.5
RORA-VLMGenerator Model=LLaVA-1.5-7B, Gen. FT=false, Ret. FT=true2025.05—27.3425.1
RORA-VLMGenerator=LLaVA-v1.5-7B, Retriever=CLIP ViT-L/142025.11—25.127.3