Video understanding on MVBench (val)
54.9AccuracyLLaVA-NeXT-Interleave-14B
Evaluation Results
| Method | Links | |
|---|---|---|
| LLaVA-NeXT-Interleave-14BModel Scale=>14B, Zero-shot transfer=true2024.09 | 54.9 | |
| MM1.5-30BModel Scale=>14B, Zero-shot transfer=true2024.09 | 54 | |
| LLaVA-NeXT-Interleave-7BModel Scale=7B, Zero-shot transfer=true2024.09 | 53.1 | |
| Mantis-Idefics2-8BModel Scale=7B, Zero-shot transfer=true2024.09 | 51.4 | |
| MM1.5-3B-MoEModel Scale=3B, Zero-shot transfer=true2024.09 | 50.3 | |
| MM1.5-1B-MoEModel Scale=1B, Zero-shot transfer=true2024.09 | 48.3 | |
| MM1.5-7BModel Scale=7B, Zero-shot transfer=true2024.09 | 48.3 | |
| MM1.5-3BModel Scale=3B, Zero-shot transfer=true2024.09 | 47.7 | |
| MM1-30BModel Scale=>14B, Zero-shot transfer=true2024.09 | 47.1 | |
| Phi-3-Vision-4BModel Scale=3B, Zero-shot transfer=true2024.09 | 46.7 | |
| MM1.5-1BModel Scale=1B, Zero-shot transfer=true2024.09 | 45.8 | |
| LLaVA-NeXT-Interleave-0.5BModel Scale=1B, Zero-shot transfer=true2024.09 | 45.6 | |
| LLaVAOneVision-0.5BModel Scale=1B, Zero-shot transfer=true2024.09 | 45.5 | |
| MM1-7BModel Scale=7B, Zero-shot transfer=true2024.09 | 45.3 | |
| MM1-3BModel Scale=3B, Zero-shot transfer=true2024.09 | 44.8 | |
| MM1-1BModel Scale=1B, Zero-shot transfer=true2024.09 | 43.8 | |
| GPT-4VModel Scale=>14B, Zero-shot transfer=true2024.09 | 43.5 | |
| Emu2-Chat-37BModel Scale=>14B, Zero-shot transfer=true2024.09 | 39.7 | |
| LLaVA-v1.5-7BModel Scale=7B, Zero-shot transfer=true2024.09 | 36 | |
| Idefics2-8BModel Scale=7B, Zero-shot transfer=true2024.09 | 29.7 |