MetaAug: Meta-Data Augmentation for Post-Training Quantization
Cuong Pham, Hoang Anh Dung, Cuong C. Nguyen, Trung Le, Dinh Phung, Gustavo Carneiro, Thanh-Toan Do
TL;DR
This paper tackles overfitting in post-training quantization by introducing MetaAug, a meta-learning framework that jointly optimizes a data-transformer T and a quantized model using bi-level optimization. The transformation T augments calibration data to train the quantized network while validation occurs on the original calibration set, with information-preserving and margin regularizations ensuring the transformation remains informative yet non-trivial. Experimental results on ImageNet across ResNet-18, ResNet-50, and MobileNetV2 show MetaAug consistently outperforms state-of-the-art PTQ methods and reduces the train-test accuracy gap, validating the approach’s effectiveness. The work also analyzes various information-preservation losses, demonstrates the benefits of distribution-preserving metrics, and discusses augmentation interactions, establishing MetaAug as a practical improvement for PTQ with limited calibration data.
Abstract
Post-Training Quantization (PTQ) has received significant attention because it requires only a small set of calibration data to quantize a full-precision model, which is more practical in real-world applications in which full access to a large training set is not available. However, it often leads to overfitting on the small calibration dataset. Several methods have been proposed to address this issue, yet they still rely on only the calibration set for the quantization and they do not validate the quantized model due to the lack of a validation set. In this work, we propose a novel meta-learning based approach to enhance the performance of post-training quantization. Specifically, to mitigate the overfitting problem, instead of only training the quantized model using the original calibration set without any validation during the learning process as in previous PTQ works, in our approach, we both train and validate the quantized model using two different sets of images. In particular, we propose a meta-learning based approach to jointly optimize a transformation network and a quantized model through bi-level optimization. The transformation network modifies the original calibration data and the modified data will be used as the training set to learn the quantized model with the objective that the quantized model achieves a good performance on the original calibration data. Extensive experiments on the widely used ImageNet dataset with different neural network architectures demonstrate that our approach outperforms the state-of-the-art PTQ methods.