OFA-X-MT

Reported on 6 benchmarks across 3 tasks · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Natural Language Processing6 results

Visual Question Answering (VQA)onVQA-X
Accuracy· 2022-12-08
92.6
SOTA
Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations arXiv:2212.04231
Explanation GenerationonVCR
Human Explanation Rating· 2022-12-08
77.3
SOTA
Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations arXiv:2212.04231
Visual Question Answering (VQA)onVCR (Q-A) test
Accuracy· 2022-12-08
62
best: 89.4 (GPT4RoI)
Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations arXiv:2212.04231
Natural Language Inferenceone-SNLI-VE
Accuracy· 2022-12-08
78.9
best: 80.9 (OFA-X)
Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations arXiv:2212.04231
Explanation GenerationonVQA-X
Human Explanation Rating· 2022-12-08
87.8
best: 89.5 (OFA-X)
Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations arXiv:2212.04231
Explanation Generationone-SNLI-VE
Human Explanation Rating· 2022-12-08
80.4
best: 85.7 (OFA-X)
Harnessing the Power of Multi-Task Pretraining for Ground-Truth Level Natural Language Explanations arXiv:2212.04231