List your company

Zero-Shot Evaluation on Image Task Suite (VQAv2, GQA, TextVQA, VizWiz, ScienceQA, POPE, OK-VQA)

62.4Accuracy (%)MMER
52.166454.823257.4860.1368May 21, 2025

Evaluation results

MethodLinks
MMEREvaluation Protocol=Merged2025.0562.4100.3
Original MLLMsEvaluation Protocol=Zero-shot2025.0562.23—
PCB-MergingEvaluation Protocol=Merged2025.0556.0290
TA (w/ DARE)Evaluation Protocol=Merged2025.0555.4788.4
TAEvaluation Protocol=Merged2025.0554.8987.5
TIES (w/ DARE)Evaluation Protocol=Merged2025.0554.7387.2
TIESEvaluation Protocol=Merged2025.0554.186.2
NaiveMC (w/ DARE)Evaluation Protocol=Merged2025.0553.6185.3
NaiveMCEvaluation Protocol=Merged2025.0552.5683.6