Quantization-Guided Training for Compact TinyML Models
This work addresses the need for compact models in resource-constrained TinyML applications, offering an incremental improvement over existing quantization-aware training methods.
The paper tackles the problem of compressing deep neural networks for TinyML by proposing Quantization Guided Training (QGT), which uses regularization to optimize low-bit-precision targets, resulting in an 81KB model for person detection at 2-bit precision with only a 3% accuracy drop compared to a floating-point baseline.
We propose a Quantization Guided Training (QGT) method to guide DNN training towards optimized low-bit-precision targets and reach extreme compression levels below 8-bit precision. Unlike standard quantization-aware training (QAT) approaches, QGT uses customized regularization to encourage weight values towards a distribution that maximizes accuracy while reducing quantization errors. One of the main benefits of this approach is the ability to identify compression bottlenecks. We validate QGT using state-of-the-art model architectures on vision datasets. We also demonstrate the effectiveness of QGT with an 81KB tiny model for person detection down to 2-bit precision (representing 17.7x size reduction), while maintaining an accuracy drop of only 3% compared to a floating-point baseline.