Nesting Forward Automatic Differentiation for Memory-Efficient Deep Neural Network Training
Cong Guo, Yuxian Qiu, Jingwen Leng, Chen Guang Zhang, Ying Shan Cao, Quanlu Zhang, Yunxin Liu, Fan Yang, Minyi Guo · 2022 IEEE 40th International Conference on Computer Design (ICCD) · 2022
An activation function is an element-wise mathematical function and plays a crucial role in deep neural networks (DNN). Many novel and sophisticated activation functions have been proposed to improve the DNN accuracy but also consume massive memory in the training process with back-propagation. In this study, we propose the nested forward automatic differentiation (Forward-AD), specifically for the element-wise activation function for memory-efficient DNN training. We deploy nested Forward-AD in two widely-used deep learning frameworks, TensorFlow and PyTorch, which support the static and dynamic computation graph, respectively. Our evaluation shows that nested Forward-AD reduces the memory footprint by up to 1.97× than the baseline model and outperforms the recomputation by 20% under the same memory reduction ratio.