BLT: Bidirectional Layout Transformer for Controllable Layout Generation
This work addresses scalable and diverse visual design for graphic designers, representing an incremental improvement with specific speed and control gains.
The paper tackles the problem of automatic layout generation for graphic design by introducing BLT, a bidirectional layout transformer, which achieves up to 10x speedup in inference time compared to state-of-the-art models while enabling controllable layout generation.
Creating visual layouts is a critical step in graphic design. Automatic generation of such layouts is essential for scalable and diverse visual designs. To advance conditional layout generation, we introduce BLT, a bidirectional layout transformer. BLT differs from previous work on transformers in adopting non-autoregressive transformers. In training, BLT learns to predict the masked attributes by attending to surrounding attributes in two directions. During inference, BLT first generates a draft layout from the input and then iteratively refines it into a high-quality layout by masking out low-confident attributes. The masks generated in both training and inference are controlled by a new hierarchical sampling policy. We verify the proposed model on six benchmarks of diverse design tasks. Experimental results demonstrate two benefits compared to the state-of-the-art layout transformer models. First, our model empowers layout transformers to fulfill controllable layout generation. Second, it achieves up to 10x speedup in generating a layout at inference time than the layout transformer baseline. Code is released at https://shawnkx.github.io/blt.