Collaborative Distillation For Ultra-Resolution Universal Style Transfer
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)(2020)
摘要
Universal style transfer methods typically leverage rich representations from deep Convolutional Neural Network (CNN) models (e.g., VGG-19) pre-trained on large collections of images. Despite the effectiveness, its application is heavily constrained by the large model size to handle ultra-resolution images given limited memory. In this work, we present a new knowledge distillation method (named Collaborative Distillation) for encoder-decoder based neural style transfer to reduce the convolutional filters. The main idea is underpinned by a finding that the encoder-decoder pairs construct an exclusive collaborative relationship, which is regarded as a new kind of knowledge for style transfer models. Moreover, to overcome the feature size mismatch when applying collaborative distillation, a linear embedding loss is introduced to drive the student network to learn a linear embedding of the teacher's features. Extensive experiments show the effectiveness of our method when applied to different universal style transfer approaches (WCT and AdaIN), even if the model size is reduced by 15.5 times. Especially, on WCT with the compressed models, we achieve ultra-resolution (over 40 megapixels) universal style transfer on a 12GB GPU for the first time. Further experiments on optimization-based stylization scheme show the generality of our algorithm on different stylization paradigms. Our code and trained models are available at https://github.com/mingsun-tse/collaborative-distillation.
更多查看译文
关键词
ultra-resolution images,knowledge distillation method,collaborative distillation,encoder-decoder based neural style transfer,convolutional filters,encoder-decoder pairs,exclusive collaborative relationship,style transfer models,feature size mismatch,linear embedding loss,student network,universal style transfer,compressed models,trained models,ultra-resolution universal style transfer,universal style transfer methods,leverage rich representations,deep convolutional neural network models,VGG-19
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络