Tasks SotA Datasets Papers Methods Submit About

Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable Benchmarks All SotA Datasets Papers Methods

Community

Submit Results About

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/OFA

OFA

Reported on 22 benchmarks across 6 tasks · 2 papers · 15 SOTA

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

Natural Language Processing18 results

Visual Question Answering (VQA)onGRIT
VQA (ablation)· 2022-02-07
72.4
best: 74.5 (Unified-IOXL)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question Answering (VQA)onVQA v2 test-std
number· 2022-02-07
71.44
best: 72.24 (ONE-PEACE)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question Answering (VQA)onVQA v2 test-std
other· 2022-02-07
73.35
best: 77.02 (mPLUG-Huge)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question Answering (VQA)onVQA v2 test-std
overall· 2022-02-07
81.98
best: 84.03 (BEiT-3)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Natural Language InferenceonSNLI-VE val
Accuracy· 2022-02-07
91
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Natural Language InferenceonSNLI-VE test
Accuracy· 2022-02-07
91.2
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Image CaptioningonCOCO Captions
BLEU-4· 2022-02-07
44.9
best: 46.5 (mPLUG)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Image CaptioningonCOCO Captions
CIDER· 2022-02-07
154.9
best: 155.1 (mPLUG)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Image CaptioningonCOCO Captions
SPICE· 2022-02-07
26.6
best: 27 (VAST)
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question AnsweringonVQA v2 test-dev
Accuracy· 2022-02-07
82
best: 82.3 (BLIP-2 ViT-G OPT 6.7B (fine-tuned))
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question AnsweringonGRIT
VQA (ablation)· 2022-02-07
72.4
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question AnsweringonVQA v2 test-std
number· 2022-02-07
71.44
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question AnsweringonVQA v2 test-std
other· 2022-02-07
73.35
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question AnsweringonVQA v2 test-std
overall· 2022-02-07
81.98
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question AnsweringonVQA v2 test-std
yes/no· 2022-02-07
94.66
SOTA
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question Answering (VQA)onVQA v2 test-dev
Accuracy· 2022-02-07
82
best: 84.3 (PaLI)
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Visual Question Answering (VQA)onVQA v2 test-std
yes/no· 2022-02-07
94.66
best: 94.85 (ONE-PEACE)
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Image CaptioningonCOCO Captions
METEOR· 2022-02-07
32.5
best: 33.9 (CoCa)
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052

Knowledge Base3 results

Text SummarizationonGigaWord
ROUGE-1· 2022-02-07
39.81
best: 60.12 (OpenAI/o3-mini)
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Text SummarizationonGigaWord
ROUGE-2· 2022-02-07
20.66
best: 54.22 (OpenAI/o3-mini)
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052
Text SummarizationonGigaWord
ROUGE-L· 2022-02-07
37.11
best: 60.29 (Riple/Saanvi-v0.1)
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework arXiv:2202.03052

Methodology1 result

Anomaly DetectiononUCR Anomaly Archive
AUC ROC · 2024-11-01
0.5699
best: 0.8001 (SubLOF)
KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks arXiv:2411.00278