研究人员开发了 Tuna-2，这是一种新颖的统一多模态模型，它绕过了传统的视觉编码器来进行视觉理解和生成。通过直接处理像素嵌入，Tuna-2 简化了架构，并实现了从原始像素到端到端的优化。实验表明，这种像素空间方法在多模态基准测试中取得了最先进的成果，在生成高质量图像方面优于潜在空间方法，并展现出卓越的多模态理解能力，尤其是在需要详细视觉感知的任务上。 AI
影响 消除了多模态模型中对预训练视觉编码器的需求，可能简化架构并提高性能。
排序理由 这是一篇描述新模型及其在基准测试中表现的研究论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
Unified multimodal models typically rely on pretrained vision encoders and use separate visual representations for understanding and generation, creating misalignment between the two tasks and preventing fully end-to-end optimization from raw pixels. We introduce Tuna-2, a native…
arXiv:2604.24763v1 Announce Type: new Abstract: Unified multimodal models typically rely on pretrained vision encoders and use separate visual representations for understanding and generation, creating misalignment between the two tasks and preventing fully end-to-end optimizatio…
Unified multimodal models typically rely on pretrained vision encoders and use separate visual representations for understanding and generation, creating misalignment between the two tasks and preventing fully end-to-end optimization from raw pixels. We introduce Tuna-2, a native…