← back to paper
arxiv: 2605.00809 · 2 revisions
Let ViT Speak: Generative Language-Image Pre-training