List your company
ResearchFormerly WizWand

Visual Perception on SEED-Bench Image

73.7AccuracyMG-LLaVA
45.30852.67960.0567.421Jun 25, 2024
Updated 8mo ago

Evaluation results

MethodLinks
MG-LLaVALLM=Yi1.5-34B, Params.=34.4B, Res.=336 (768)2024.0673.7
GPT-4V2024.0672.3
SPHINX-1kLLM=Vicuna-7B, Params.=10B, Res.=4482024.0671.6
MG-LLaVALLM=LLaMA3-8B, Params.=8.4B, Res.=336 (768)2024.0671.5
MG-LLaVALLM=Vicuna-13B, Params.=13.6B, Res.=336 (768)2024.0670.8
GeminiPro Vision2024.0670.7
MG-LLaVALLM=Phi3-3.8B, Params.=4.2B, Res.=336 (768)2024.0670.2
MG-LLaVALLM=Vicuna-7B, Params.=7.4B, Res.=336 (768)2024.0669.4
SPHINXLLM=Vicuna-7B, Params.=10B, Res.=2242024.0669.1
LLaVA-1.5LLM=Vicuna-13B, Params.=13.4B, Res.=3362024.0668.2
MiniCPM-V2LLM=MiniCPM-2.4B, Params.=2.8B, Res.=4482024.0667.1
LLaVA-1.5LLM=Vicuna-7B, Params.=7.2B, Res.=3362024.0666.1
Qwen-VL-Plus2024.0665.7
LLaVA-HRLLM=Vicuna-13B, Params.=13.4B, Res.=448 (1024)2024.0664.5
Qwen-VL-ChatLLM=Qwen-7B, Params.=9.6B, Res.=4482024.0658.2
Qwen-VLLLM=Qwen-7B, Params.=9.6B, Res.=4482024.0656.3
InstructBLIPLLM=Vicuna-7B, Params.=8.2B, Res.=2242024.0653.4
BLIP-2LLM=Vicuna-13B, Params.=14.2B, Res.=2242024.0646.4