Paper · arXiv 2404.01143
Condition-Aware Neural Network for Controlled Image Generation
Han Cai, Muyang Li, Zhuoyang Zhang, Qinsheng Zhang, Ming-Yu Liu, Song Han
A Condition-Aware Neural Network (CAN) method for image generation dynamically adjusts neural network weights based on input conditions, demonstrating significant improvements for diffusion transformer models and surpassing existing models in terms of FID score and computational efficiency.
Abstract
We present Condition-Aware Neural Network (CAN), a new method for adding control to image generative models. In parallel to prior conditional control methods, CAN controls the image generation process by dynamically manipulating the weight of the neural network. This is achieved by introducing a condition-aware weight generation module that generates conditional weight for convolution/linear layers based on the input condition. We test CAN on class-conditional image generation on ImageNet and text-to-image generation on COCO. CAN consistently delivers significant improvements for diffusion transformer models, including DiT and UViT. In particular, CAN combined with EfficientViT (CaT) achieves 2.78 FID on ImageNet 512x512, surpassing DiT-XL/2 while requiring 52x fewer MACs per sampling step.