Visual Question Answering on RealWorldQA 1.0 (test)
0.6353AccuracySCL
Evaluation Results
| Method | Links | |
|---|---|---|
| SCLBackbone=MiniCPM-Llama3-V2.5, Strategy=SCL, Rank=12024.10 | 0.6353 | |
| SFTBackbone=MiniCPM-Llama3-V2.5, Strategy=SFT, Rank=22024.10 | 0.6235 | |
| SCLBackbone=InternLM-XComposer-2-7B, Strategy=SCL, Rank=1.132024.10 | 0.6223 | |
| SFTBackbone=InternLM-XComposer-2-7B, Strategy=SFT, Rank=1.752024.10 | 0.6203 | |
| MiniCPM-Llama3-V2.5Backbone=MiniCPM-Llama3-V2.5, Strategy=Base, Rank=32024.10 | 0.617 | |
| InternLM-XComposer-2-7BBackbone=InternLM-XComposer-2-7B, Strategy=Base, Rank=32024.10 | 0.6045 | |
| SFTBackbone=LLaVA-V1.5-13B, Strategy=SFT, Rank=1.882024.10 | 0.5659 | |
| SCLBackbone=LLaVA-V1.5-13B, Strategy=SCL, Rank=1.382024.10 | 0.5603 | |
| LLaVA-V1.5-13BBackbone=LLaVA-V1.5-13B, Strategy=Base, Rank=2.752024.10 | 0.5485 | |
| SCLBackbone=LLaVA-V1.5-7B, Strategy=SCL, Rank=1.632024.10 | 0.5271 | |
| SFTBackbone=LLaVA-V1.5-7B, Strategy=SFT, Rank=32024.10 | 0.5221 | |
| POVIDBackbone=LLaVA-V1.5-7B, Strategy=POVID, Rank=3.252024.10 | 0.515 | |
| CSRBackbone=LLaVA-V1.5-7B, Strategy=CSR, Rank=4.252024.10 | 0.5103 | |
| LLaVA-V1.5-7BBackbone=LLaVA-V1.5-7B, Strategy=Base, Rank=4.52024.10 | 0.509 | |
| SIMABackbone=LLaVA-V1.5-7B, Strategy=SIMA, Rank=4.382024.10 | 0.4941 |