Efficient and Robust Quantization-aware Training via Adaptive Coreset Selection
This work addresses the high computational cost and robustness issues in model compression for deep learning practitioners, offering an incremental improvement over existing QAT methods.
The paper tackles the inefficiency and vulnerability to label noise in quantization-aware training (QAT) by proposing an adaptive coreset selection method based on error vector and disagreement scores, achieving a 68.39% accuracy with 4-bit quantized ResNet-18 on ImageNet-1K using only 10% of data, a 4.24% absolute gain over the baseline.
Quantization-aware training (QAT) is a representative model compression method to reduce redundancy in weights and activations. However, most existing QAT methods require end-to-end training on the entire dataset, which suffers from long training time and high energy costs. In addition, the potential label noise in the training data undermines the robustness of QAT. We propose two metrics based on analysis of loss and gradient of quantized weights: error vector score and disagreement score, to quantify the importance of each sample during training. Guided by these two metrics, we proposed a quantization-aware Adaptive Coreset Selection (ACS) method to select the data for the current training epoch. We evaluate our method on various networks (ResNet-18, MobileNetV2, RetinaNet), datasets(CIFAR-10, CIFAR-100, ImageNet-1K, COCO), and under different quantization settings. Specifically, our method can achieve an accuracy of 68.39\% of 4-bit quantized ResNet-18 on the ImageNet-1K dataset with only a 10\% subset, which has an absolute gain of 4.24\% compared to the baseline. Our method can also improve the robustness of QAT by removing noisy samples in the training set.