Re-parameterizing Your Optimizers rather than Architectures
This addresses the issue of using model-agnostic optimizers for neural networks, offering a practical solution for improving training and inference efficiency in machine learning applications.
The paper tackles the problem of incorporating model-specific prior knowledge into optimizers by proposing Gradient Re-parameterization, resulting in RepOptimizers that enable a simple VGG-style model (RepOpt-VGG) to perform on par with or better than recent well-designed models, with high inference speed and training efficiency.
The well-designed structures in neural networks reflect the prior knowledge incorporated into the models. However, though different models have various priors, we are used to training them with model-agnostic optimizers such as SGD. In this paper, we propose to incorporate model-specific prior knowledge into optimizers by modifying the gradients according to a set of model-specific hyper-parameters. Such a methodology is referred to as Gradient Re-parameterization, and the optimizers are named RepOptimizers. For the extreme simplicity of model structure, we focus on a VGG-style plain model and showcase that such a simple model trained with a RepOptimizer, which is referred to as RepOpt-VGG, performs on par with or better than the recent well-designed models. From a practical perspective, RepOpt-VGG is a favorable base model because of its simple structure, high inference speed and training efficiency. Compared to Structural Re-parameterization, which adds priors into models via constructing extra training-time structures, RepOptimizers require no extra forward/backward computations and solve the problem of quantization. We hope to spark further research beyond the realms of model structure design. Code and models \url{https://github.com/DingXiaoH/RepOptimizers}.