List your company
ResearchFormerly WizWand

Multi-choice Visual Question Answering on A-OKVQA

82.71AccuracyCVLM
43.564453.727263.8974.0528Nov 23, 2023Mar 25, 2024Jul 27, 2024Nov 28, 2024Apr 1, 2025Aug 3, 2025Dec 5, 2025
Updated 8mo ago

Evaluation results

MethodLinks
CVLMLLM=Qwen-VL2024.0282.71
BaselineModel=LLaVA-13B, TFLOPS=5.81, FLOPs Ratio=100%2025.0382
HiMAPModel=LLaVA-13B, TFLOPS=1.36, FLOPs Ratio=23%2025.0381.4
FastVModel=LLaVA-13B, TFLOPS=3.09, FLOPs Ratio=53%2025.0381.3
HiMAPModel=InternVL-7B, TFLOPS=0.56, FLOPs Ratio=20%2025.0380.1
BaselineModel=InternVL-7B, TFLOPS=2.71, FLOPs Ratio=100%2025.0379.6
CVLM (3M IKPairs, Objects=3)LLM=Vicuna-7B, Objects=32024.0279.21
FastVModel=InternVL-7B, TFLOPS=1.39, FLOPs Ratio=52%2025.0379.1
CVLM (3M IKPairs, Objects=5)LLM=Vicuna-7B, Objects=52024.0278.95
CVLM (3M IKPairs, Objects=8)LLM=Vicuna-7B, Objects=82024.0278.95
CVLM (3M IKPairs, Objects=1)LLM=Vicuna-7B, Objects=12024.0278.69
InstructBLIPLLM=Flan-T5-XXL2024.0278.34
CVLM (3M IKPairs) w/o FKALLM=Vicuna-7B2024.0277.9
CVLMLLM=Vicuna-7B2024.0277.64
HiMAPModel=LLaVA-7B, TFLOPS=0.73, FLOPs Ratio=24%2025.0377.2
CVLM w/o FKALLM=Vicuna-7B2024.0277.12
FastVModel=LLaVA-7B, TFLOPS=1.56, FLOPs Ratio=54%2025.0377
InstructBLIPLLM=Flan-T5-XL2024.0276.68
BaselineModel=LLaVA-7B, TFLOPS=2.98, FLOPs Ratio=100%2025.0376.6
HiMAPModel=QwenVL-7B, TFLOPS=0.89, FLOPs Ratio=25%2025.0375.9
BaselineModel=QwenVL-7B, TFLOPS=3.6, FLOPs Ratio=100%2025.0375.7
FastVModel=QwenVL-7B, TFLOPS=1.9, FLOPs Ratio=53%2025.0375.3
CVLM w/o (FKA & VKA)LLM=Vicuna-7B2024.0273.97
LLaVA-v1.5†LLM=Vicuna-7B2024.0273.45
MC-CoT_BaseVision Model=DETR [11], Text Model=UnifiedQABase [24], Parameters=223M2023.1171
MINDbaseLearning=Fine-tune2025.1270.6
BLIP-2Vision Model=CLIP-VIT-LARGE [41], Text Model=FlanT5XXL [15], Parameters=11B2023.1170.2
Qwen-VLLLM=Qwen2024.0270.04
GPV-2Vision Model=VinVL [59], Text Model=T5-Base [42], Parameters=300M2023.1160.3
GPV-2Learning=Fine-tune2025.1260.3
IPVR (GPT-3)Learning=Few-shot2025.1258.7
ClipCapLearning=Few-shot2025.1256.9
KRISPVision Model=Faster R-CNN [44], Text Model=BERT [23], Parameters=200M2023.1151.9
KRISPLearning=Fine-tune2025.1251.9
LXMERTVision Model=Transformer [49], Text Model=Transformer [49], Parameters=220M2023.1151.4
LXMERTLearning=Fine-tune2025.1251.4
Multimodal-CoT_BaseVision Model=DETR [11], Text Model=UnifiedQABase [24], Parameters=223M2023.1150.6
Multimodal-CoTbaseLearning=Fine-tune2025.1250.6
VILBERTVision Model=Faster R-CNN [44], Text Model=BERT [23], Parameters=300M2023.1149.1
ViLBERTLearning=Fine-tune2025.1249.1
PythiaVision Model=ResNet [19], Text Model=BERT [23], Parameters=70M2023.1149
PythiaLearning=Fine-tune2025.1249
IPVRVision Model=Faster-RCNN [44], Text Model=OPT [62], Parameters=66B2023.1148.6
IPVR (OPT-66B)Learning=Few-shot2025.1248.6
PaLM-COTText Model=PaLM [13], Parameters=540B2023.1148.1
CoTLearning=Few-shot2025.1248.1
PICaVision Model=VinVL [59], Text Model=GPT-3 [10], Parameters=175B2023.1146.1
PicaLearning=Few-shot2025.1246.1
InstructBLIPLLM=Vicuna-7B2024.0245.07