Visual Instruction Following on LLaVA-W
102ScoreGPT-4o
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4oZero-shot=true, Decoding Strategy=greedy2024.09 | 102 | |
| Gemini-1.5-ProZero-shot=true, Decoding Strategy=greedy2024.09 | 95.3 | |
| GPT-4VZero-shot=true, Decoding Strategy=greedy2024.09 | 93.1 | |
| TroLModel size=7B2024.06 | 92.8 | |
| LLaVA-NeXT-34BModel Scale=30B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 89.6 | |
| MM1-30BModel Scale=30B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 89.3 | |
| LLaVA-NeXT-7BModel Scale=7B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 81.6 | |
| MM1-7BModel Scale=7B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 81.5 | |
| MM1.5-30BModel Scale=30B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 80.4 | |
| DeepSeek-VLModel size=7B2024.06 | 78.9 | |
| TroLModel size=3.8B2024.06 | 76.6 | |
| MM1.5-3B-MoEModel Scale=3B, Architecture=MoE, Zero-shot=true, Decoding Strategy=greedy2024.09 | 76.1 | |
| TinyLLaVAModel Scale=3B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 75.8 | |
| MM1.5-1B-MoEModel Scale=1B, Architecture=MoE, Zero-shot=true, Decoding Strategy=greedy2024.09 | 75.5 | |
| MM1.5-7BModel Scale=7B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 74.2 | |
| MM1.5-3BModel Scale=3B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 73 | |
| MM1-3BModel Scale=3B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 72.1 | |
| CoLLaVOModel size=7B2024.06 | 71.9 | |
| MM1.5-1BModel Scale=1B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 71.6 | |
| Phi-3-Vision-4BModel Scale=3B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 71.6 | |
| MoAIModel size=7B2024.06 | 69.7 | |
| LAFModel size=7B2024.06 | 69.5 | |
| MiniCPM-V 2.0-3BModel Scale=3B, Zero-shot=true, Decoding Strategy=beam search2024.09 | 69.2 | |
| MM1-1BModel Scale=1B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 67.5 | |
| TinyLLaVAModel Scale=1B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 60.8 | |
| InternVL2-2BModel Scale=1B, Zero-shot=true, Decoding Strategy=beam search2024.09 | 60 | |
| SPHINX-TinyModel Scale=1B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 52.3 | |
| LLaVAOne Vision-0.5BModel Scale=1B, Zero-shot=true, Decoding Strategy=greedy2024.09 | 29.1 |