Mobile V-MoEs: Scaling Down Vision Transformers via Sparse Mixture-of-Experts
This work addresses efficiency challenges for deploying vision models on mobile devices, representing an incremental advance in adapting existing methods to new constraints.
The paper tackles the problem of scaling down Vision Transformers for resource-constrained applications by using sparse Mixture-of-Experts, achieving improvements such as 3.39% on ImageNet-1k for ViT-Tiny and 4.66% for a smaller variant with 54M FLOPs.
Sparse Mixture-of-Experts models (MoEs) have recently gained popularity due to their ability to decouple model size from inference efficiency by only activating a small subset of the model parameters for any given input token. As such, sparse MoEs have enabled unprecedented scalability, resulting in tremendous successes across domains such as natural language processing and computer vision. In this work, we instead explore the use of sparse MoEs to scale-down Vision Transformers (ViTs) to make them more attractive for resource-constrained vision applications. To this end, we propose a simplified and mobile-friendly MoE design where entire images rather than individual patches are routed to the experts. We also propose a stable MoE training procedure that uses super-class information to guide the router. We empirically show that our sparse Mobile Vision MoEs (V-MoEs) can achieve a better trade-off between performance and efficiency than the corresponding dense ViTs. For example, for the ViT-Tiny model, our Mobile V-MoE outperforms its dense counterpart by 3.39% on ImageNet-1k. For an even smaller ViT variant with only 54M FLOPs inference cost, our MoE achieves an improvement of 4.66%.