BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstruction

TL;DR

BRECQ is a novel post-training quantization method that achieves INT2 bitwidth, enabling faster and comparable performance to quantization-aware training for neural networks.

Contribution

It introduces a new PTQ framework that pushes bitwidth down to INT2 and incorporates mixed precision, with theoretical analysis and extensive experiments.

Findings

Achieves INT2 quantization for neural networks.

Attains 4-bit performance comparable to QAT.

Speeds up model production by 240 times.

Abstract

We study the challenging task of neural network quantization without end-to-end retraining, called Post-training Quantization (PTQ). PTQ usually requires a small subset of training data but produces less powerful quantized models than Quantization-Aware Training (QAT). In this work, we propose a novel PTQ framework, dubbed BRECQ, which pushes the limits of bitwidth in PTQ down to INT2 for the first time. BRECQ leverages the basic building blocks in neural networks and reconstructs them one-by-one. In a comprehensive theoretical study of the second-order error, we show that BRECQ achieves a good balance between cross-layer dependency and generalization error. To further employ the power of quantization, the mixed precision technique is incorporated in our framework by approximating the inter-layer and intra-layer sensitivity. Extensive experiments on various handcrafted and searched neural architectures are conducted for both image classification and object detection tasks. And for the first time we prove that, without bells and whistles, PTQ can attain 4-bit ResNet and MobileNetV2 comparable with QAT and enjoy 240 times faster production of quantized models. Codes are available at https://github.com/yhhhli/BRECQ.