Multi-modal Large Language Model Evaluation on MMBench (test)
64.4MMBench ScoreOPERA
Evaluation results
| Method | Links | |
|---|---|---|
| OPERADecoding Strategy=OPERA, Model=LLaVA-1.5 7B2023.11 | 64.4 | |
| GreedyDecoding Strategy=Greedy, Model=LLaVA-1.5 7B2023.11 | 64.3 | |
| BeamDecoding Strategy=Beam Search, Model=LLaVA-1.5 7B2023.11 | 64.1 | |
| NucleusDecoding Strategy=Nucleus Sampling, Model=LLaVA-1.5 7B2023.11 | 64 | |
| DoLaDecoding Strategy=DoLa, Model=LLaVA-1.5 7B2023.11 | 63.8 |