List your company
ResearchFormerly WizWand

PET on Different-environment out-of-domain

87.7AccuracyBagel + Mixed Training
42.9854.5966.277.81Jun 2, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Bagel + Mixed TrainingModel Category=Ours (fine-tuned BAGEL), Evaluation Protocol=Fine-tuned, Reasoning Strategy=IPT + Label-only mixture2026.0687.7
Bagel + IPTModel Category=Ours (fine-tuned BAGEL), Evaluation Protocol=Fine-tuned, Reasoning Strategy=Imaginative Perception Token2026.0687
Bagel (label-only)Model Category=Ours (fine-tuned BAGEL), Evaluation Protocol=Fine-tuned, Reasoning Strategy=Answer supervision only2026.0682
Bagel + Text CoTModel Category=Ours (fine-tuned BAGEL), Evaluation Protocol=Fine-tuned, Reasoning Strategy=Textual chain-of-thought2026.0670.3
GPT-5Model Category=VQA Models, Evaluation Protocol=Zero-shot2026.0669.3
Gemini 2.5 FlashModel Category=VQA Models, Evaluation Protocol=Zero-shot2026.0666.3
Bagel (base)Model Category=Ours (fine-tuned BAGEL), Evaluation Protocol=Zero-shot2026.0662.7
Qwen2.5-VL-7BModel Category=VQA Models, Evaluation Protocol=Zero-shot2026.0654.3
GPT-5.2Model Category=VQA Models, Evaluation Protocol=Zero-shot2026.0654
Gemini 3 FlashModel Category=VQA Models, Evaluation Protocol=Zero-shot2026.0651.3
InternVL3.5-8BModel Category=VQA Models, Evaluation Protocol=Zero-shot2026.0647.7
Chameleon 7BModel Category=Unified Models, Evaluation Protocol=Zero-shot2026.0647.3
Qwen3-VL-8BModel Category=VQA Models, Evaluation Protocol=Zero-shot2026.0646.7
Janus-Pro-7BModel Category=Unified Models, Evaluation Protocol=Zero-shot2026.0644.7