Research map / Vision-language and multimodal / Cross-modal learning
OFA: Unifying Architectures, Tasks, and Modalities Through a Simple Sequence-to-Sequence Learning Framework
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Scale OFA from Base to Huge | OFA · VQAv2 test-std | accuracy: 78.1 -> 82.0 | Helped |
| Scale OFA from Base to Huge | OFA · SNLI-VE test | accuracy: 89.2 -> 91.2 | Helped |
| Add text infilling pretraining (ablation) | OFA Base · Image captioning | CIDEr: +0.8 CIDEr | Helped |
| Add text infilling pretraining (ablation) | OFA Base · VQA | accuracy: +0.46 Acc. | Helped |
| Add text infilling pretraining (ablation) | OFA Base · ImageNet-1K | accuracy: -1.0 Acc. | Hurt |
| Add text infilling pretraining (ablation) | OFA Base · Text-to-image generation | CLIPSIM: -0.1 CLIPSIM | No clear effect |
| Add image infilling pretraining (ablation) | OFA Base · ImageNet-1K | accuracy: +1.0 Acc. | Helped |
| Add image infilling pretraining (ablation) | OFA Base · Text-to-image generation | CLIPSIM: +0.6 CLIPSIM | Helped |
| Add image infilling pretraining (ablation) | OFA Base · Image captioning | CIDEr: -0.7 CIDEr | Hurt |
| Add image infilling pretraining (ablation) | OFA Base · VQA | accuracy: -0.3 Acc. | Hurt |
| Add object detection pretraining (ablation) | OFA Base · Image captioning | CIDEr: +2.3 CIDEr | Helped |
| Add visual grounding and grounded captioning (ablation) | OFA Base · Image captioning | CIDEr: +1.4 CIDEr | Helped |
Closest papers by meaning
- Instruct-Imagen: Image Generation with Multi-modal Instruction (CVPR 2024 · similarity 0.88)
- MoDA: Modulation Adapter for Fine-Grained Visual Understanding in Instructional MLLMs (ICML 2026 · similarity 0.88)
- 4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities (NeurIPS 2024 · similarity 0.87)
- Vision-Language Models Create Cross-Modal Task Representations (ICML 2025 · similarity 0.87)
- Post-pre-training for Modality Alignment in Vision-Language Foundation Models (CVPR 2025 · similarity 0.87)
- LoMo: Local Modality Substitution for Deeper Vision-Language Fusion (NeurIPS 2026 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.