Image Quality Assessment on QBench
77.5AccuracyQwen2-VL-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2-VL-7BParameters=7B2024.12 | 77.5 | |
| VLsI-7BParameters=7B2024.12 | 77.5 | |
| Latent DenoisingArchitecture=Qwen-2.5-VL2026.04 | 75.9 | |
| GPT-4V2024.04 | 74.5 | |
| BaselineArchitecture=Qwen-2.5-VL2026.04 | 73.9 | |
| Phantom-7BParameters=7B2024.12 | 73.8 | |
| TroL-7BParameters=7B2024.12 | 73.6 | |
| TroL-7BParameters=7B2024.06 | 73.6 | |
| Qwen-VL-Max2024.04 | 73.6 | |
| MM1-7BParameters=7B2024.06 | 72.6 | |
| MoVALLM=Llama3-8B, PT=15M, SFT=1.6M, #IMG=5762024.04 | 70.8 | |
| MoVALLM=Hermes-Yi-34B, PT=15M, SFT=1.6M, #IMG=5762024.04 | 70.7 | |
| MoAIZero-shot=true, Number of Parameters=7B2024.03 | 70.2 | |
| LLaVA-NeXT-7BParameters=7B2024.06 | 70.1 | |
| MoVALLM=Vicuna-7B, PT=15M, SFT=1.6M, #IMG=5762024.04 | 69.3 | |
| Gemini-Pro2024.04 | 68.2 | |
| CoLLaVONumber of Parameters=7B, Setting=Zero-shot2024.02 | 67.6 | |
| SPHINX-MoE-7Bx8Architecture=MoE, Parameters=7Bx82024.06 | 66.2 | |
| SPHINX-Plus-13BParameters=13B2024.06 | 66.2 | |
| Qwen-VL-Plus2024.04 | 66 | |
| Intern-XCNumber of Parameters=7B, Setting=Zero-shot2024.02 | 64.4 | |
| Intern-XCZero-shot=true, Number of Parameters=7B2024.03 | 64.4 | |
| InternLM-XC-7BParameters=7B2024.06 | 64.4 | |
| ShareGPT4VNumber of Parameters=7B, Setting=Zero-shot2024.02 | 63.4 | |
| ShareGPT4VZero-shot=true, Number of Parameters=7B2024.03 | 63.4 | |
| ShareGPT4V-7BParameters=7B2024.06 | 63.4 | |
| mPLUG-Owl2Number of Parameters=7B, Setting=Zero-shot2024.02 | 62.9 | |
| mPLUG-Owl2Zero-shot=true, Number of Parameters=7B2024.03 | 62.9 | |
| mPLUG-Owl2-7BParameters=7B2024.06 | 62.9 | |
| LLaVA1.5Number of Parameters=13B, Setting=Zero-shot2024.02 | 62.1 | |
| LLaVA1.5Zero-shot=true, Number of Parameters=13B2024.03 | 62.1 | |
| LLaVA-1.5LLM=Vicuna-13B, PT=558K, SFT=665K, #IMG=5762024.04 | 62.1 | |
| LLaVA1.5-13BParameters=13B2024.06 | 61.4 | |
| BLIP2-13BParameters=13B2024.06 | 61 | |
| LLaVA1.5-7BParameters=7B2024.06 | 60.1 | |
| Latent DenoisingArchitecture=LLaVA+CLIP2026.04 | 59.6 | |
| Qwen-VLNumber of Parameters=7B, Setting=Zero-shot2024.02 | 59.4 | |
| Qwen-VLZero-shot=true, Number of Parameters=7B2024.03 | 59.4 | |
| Qwen-VL-7BParameters=7B2024.06 | 59.4 | |
| Qwen-VLLLM=Qwen-7B, PT=1.4B, SFT=50M, #IMG=2562024.04 | 59.4 | |
| Latent DenoisingArchitecture=LLaVA+SigLIP2026.04 | 59.1 | |
| mPLUG-OwlNumber of Parameters=7B, Setting=Zero-shot2024.02 | 58.9 | |
| mPLUG-OwlZero-shot=true, Number of Parameters=7B2024.03 | 58.9 | |
| mPLUG-Owl-7BParameters=7B2024.06 | 58.9 | |
| LLaVA1.5Number of Parameters=7B, Setting=Zero-shot2024.02 | 58.7 | |
| LLaVA1.5Zero-shot=true, Number of Parameters=7B2024.03 | 58.7 | |
| LLaVA-1.5LLM=Vicuna-7B, PT=558K, SFT=665K, #IMG=5762024.04 | 58.7 | |
| BaselineArchitecture=LLaVA+CLIP2026.04 | 58.7 | |
| Vanilla (LLaVA-1.5-7B)Retained Tokens=5762026.01 | 58.5 | |
| FastVRetained Tokens=1922026.01 | 58.1 | |
| PyramidDropRetained Tokens=1922026.01 | 58.1 | |
| PyramidDropRetained Tokens=1282026.01 | 58.1 | |
| ViTCoPRetained Tokens=1922026.01 | 57.9 | |
| FastVRetained Tokens=1282026.01 | 57.9 | |
| ViTCoPRetained Tokens=1282026.01 | 57.7 | |
| SparseVLMRetained Tokens=1922026.01 | 57.5 | |
| VisionZipRetained Tokens=1922026.01 | 57.5 | |
| BaselineArchitecture=LLaVA+SigLIP2026.04 | 57.4 | |
| SparseVLMRetained Tokens=1282026.01 | 57.2 | |
| VisionZipRetained Tokens=1282026.01 | 57 | |
| ViTCoPRetained Tokens=642026.01 | 56.8 | |
| InstructBLIPNumber of Parameters=7B, Setting=Zero-shot2024.02 | 56.7 | |
| InstructBLIP-7BParameters=7B2024.06 | 56.7 | |
| SparseVLMRetained Tokens=642026.01 | 56.3 | |
| VisionZipRetained Tokens=642026.01 | 55.9 | |
| PyramidDropRetained Tokens=642026.01 | 55.1 | |
| ShikraNumber of Parameters=13B, Setting=Zero-shot2024.02 | 54.7 | |
| ShikraZero-shot=true, Number of Parameters=13B2024.03 | 54.7 | |
| FastVRetained Tokens=642026.01 | 54 | |
| MiniGPT-4-7BParameters=7B2024.06 | 51.8 | |
| IDEFICS-9BParameters=9B2024.06 | 51.5 | |
| OtterNumber of Parameters=7B, Setting=Zero-shot2024.02 | 47.2 | |
| OtterZero-shot=true, Number of Parameters=7B2024.03 | 47.2 | |
| Otter-7BParameters=7B2024.06 | 47.2 | |
| Qwen-VL-Chat-7BParameters=7B, Variant=Chat2024.06 | 33.8 | |
| VLSI-2BParameters=2B2024.12 | 0.723 | |
| Qwen2-VL-2BParameters=2B2024.12 | 0.708 | |
| Phantom-3.8BParameters=3.8B2024.12 | 0.703 | |
| TroL-3.8BParameters=3.8B2024.12 | 0.7 | |
| Phantom-1.8BParameters=1.8B2024.12 | 0.691 | |
| TroL-1.8BParameters=1.8B2024.12 | 0.682 |