List your company

Direct Answer Visual Question Answering on A-OKVQA (test)

69AccuracyHinD-CoT-Know
13.67228.03642.456.764Nov 15, 2022May 16, 2023Nov 15, 2023May 15, 2024Nov 14, 2024May 15, 2025Nov 14, 2025

Evaluation results

MethodLinks
HinD-CoT-KnowModel Size=7B2025.1169
HinD-KnowModel Size=7B2025.1167.8
ReAuSE2025.1167.7
QACap2025.1166.3
LLaVA-1.5Model Size=7B2025.1163.7
InstructBLIPModel Size=7B, Training=Fine-tuned (FT)2025.1162.4
Qwen2-VLModel Size=7B, Training=Fine-tuned (FT)2025.1161.5
Minigpt4-v2Model Size=7B, Training=Fine-tuned (FT)2025.1161.3
PromptCap + GPT-3Example retrieval=CLIP (ViT-L/14)2022.1159.6
SimKBVQA2025.1158.6
Prophet2025.1158.2
PromptCap2025.1156.3
REVEAL2025.1152.2
MM-CoT2025.1150.6
VisualCoT2025.1150.5
GPT-22025.1148.6
GPV-22022.1140.7
KRISP2022.1127.1
ViLBERT2022.1125.9
LXMERT2022.1125.9
Pythia2022.1121.9
ClipCap2022.1115.8