Searching the Search Space of Vision Transformer
This work addresses the challenge of manually designing Vision Transformers for researchers and practitioners, offering an automated approach that yields improved performance, though it is incremental as it builds on existing neural architecture search methods.
The paper tackles the problem of automating the design of Vision Transformer architectures by using neural architecture search to evolve both the architecture and the search space, resulting in models that outperform existing ones like Swin, DeiT, and ViT on ImageNet and show effectiveness in downstream tasks such as object detection and semantic segmentation.
Vision Transformer has shown great visual representation power in substantial vision tasks such as recognition and detection, and thus been attracting fast-growing efforts on manually designing more effective architectures. In this paper, we propose to use neural architecture search to automate this process, by searching not only the architecture but also the search space. The central idea is to gradually evolve different search dimensions guided by their E-T Error computed using a weight-sharing supernet. Moreover, we provide design guidelines of general vision transformers with extensive analysis according to the space searching process, which could promote the understanding of vision transformer. Remarkably, the searched models, named S3 (short for Searching the Search Space), from the searched space achieve superior performance to recently proposed models, such as Swin, DeiT and ViT, when evaluated on ImageNet. The effectiveness of S3 is also illustrated on object detection, semantic segmentation and visual question answering, demonstrating its generality to downstream vision and vision-language tasks. Code and models will be available at https://github.com/microsoft/Cream.