List your company

General Visual Question Answering on VQA v2 (Accuracy)

83.5AccuracyFP16 (Baseline)
35.34847.84960.3572.851Sep 28, 2025Nov 2, 2025Dec 8, 2025Jan 13, 2026Feb 18, 2026Mar 26, 2026May 1, 2026

Evaluation results

MethodLinks
FP16 (Baseline)Backbone=Qwen 2.5 VL 7B, Avg. Bit width=162025.0983.5
OmniQuantBackbone=Qwen 2.5 VL 7B, Avg. Bit width=42025.0982.9
AWQBackbone=Qwen 2.5 VL 7B, Avg. Bit width=42025.0982.7
GPTQBackbone=Qwen 2.5 VL 7B, Avg. Bit width=42025.0982.6
OmniQuantBackbone=Qwen 2.5 VL 7B, Avg. Bit width=32025.0981.6
GPTQBackbone=Qwen 2.5 VL 7B, Avg. Bit width=32025.0981.5
LUQBackbone=Qwen 2.5 VL 7B, Avg. Bit width=2.75, Quantized Layers=12 layer2025.0979.7
FP16 (Baseline)Backbone=LLaVA-1.5 7B, Avg. Bit width=162025.0978.5
SlimLLMBackbone=LLaVA-1.5 7B, Avg. Bit width=42025.0976.8
AWQBackbone=LLaVA-1.5 7B, Avg. Bit width=42025.0976.6
OmniQuantBackbone=LLaVA-1.5 7B, Avg. Bit width=42025.0976.6
GPTQBackbone=LLaVA-1.5 7B, Avg. Bit width=42025.0976.3
LUQBackbone=LLaVA-1.5 7B, Avg. Bit width=2.54, Quantized Layers=16 layer2025.0974.9
SlimLLMBackbone=LLaVA-1.5 7B, Avg. Bit width=32025.0974.2
GPTQBackbone=LLaVA-1.5 7B, Avg. Bit width=32025.0973.5
OmniQuantBackbone=LLaVA-1.5 7B, Avg. Bit width=32025.0972.9
SigLIP2Arch=g/16, Data=40.0B2026.0550.1
GenLIPArch=g/16, Data=8.0B2026.0549.1
AIMv2Arch=L/14, Data=12.0B2026.0548.1
SigLIP2Arch=So/16, Data=40.0B2026.0546.5
GenLIPArch=So/16, Data=8.0B2026.0546.5
CLIPArch=L/14, Data=12.8B2026.0546.1
GenLIPArch=L/16, Data=8.0B2026.0544.4
OVision2Arch=L/16, Data=12.8B2026.0544
SigLIP2Arch=L/16, Data=40.0B2026.0543.1
SigLIPArch=L/16, Data=40.0B2026.0541.3
BiLLMBackbone=Qwen 2.5 VL 7B, Avg. Bit width=1.082025.0939.5
BiLLMBackbone=LLaVA-1.5 7B, Avg. Bit width=1.082025.0937.2