Cognition and Reasoning on GQA
0.6226ScoreTask Arithmetic
Evaluation results
| Method | Links | |
|---|---|---|
| Task Arithmeticmerging_target=Qwen2-VL-7B, merging_source=LLaVA-OneVision-7B2025.03 | 0.6226 | |
| AdaMMSmerging_target=Qwen2-VL-7B, merging_source=LLaVA-OneVision-7B2025.03 | 0.6202 | |
| MetaGPTmerging_target=Qwen2-VL-7B, merging_source=LLaVA-OneVision-7B2025.03 | 0.6199 | |
| LLaVA-v1.5-7BStatus=Source Model2025.03 | 0.6194 | |
| Ties-Mergingmerging_target=Qwen2-VL-7B, merging_source=LLaVA-OneVision-7B2025.03 | 0.6184 | |
| Qwen2-VL (base)model_status=Original Model2025.03 | 0.618 | |
| AdaMMSMerging Protocol=AdaMMS2025.03 | 0.6012 | |
| LLaVA-OneVisionmodel_status=Original Model2025.03 | 0.5984 | |
| MetaGPTMerging Protocol=MetaGPT2025.03 | 0.5944 | |
| CogVLM-chat-7BStatus=Base Model2025.03 | 0.5943 | |
| Task ArithmeticMerging Protocol=Task Arithmetic2025.03 | 0.5882 | |
| DARE-Tiesmerging_target=Qwen2-VL-7B, merging_source=LLaVA-OneVision-7B2025.03 | 0.5851 | |
| DARE-LinearMerging Protocol=DARE-Linear2025.03 | 0.5801 | |
| DARE-Linearmerging_target=Qwen2-VL-7B, merging_source=LLaVA-OneVision-7B2025.03 | 0.5513 | |
| Ties-MergingMerging Protocol=Ties-Merging2025.03 | 0.5197 | |
| DARE-TiesMerging Protocol=DARE-Ties2025.03 | 0.4962 |