Paper ID: 2306.03076

Sensitivity-Aware Finetuning for Accuracy Recovery on Deep Learning Hardware

Lakshmi Nair, Darius Bunandar

Existing methods to recover model accuracy on analog-digital hardware in the presence of quantization and analog noise include noise-injection training. However, it can be slow in practice, incurring high computational costs, even when starting from pretrained models. We introduce the Sensitivity-Aware Finetuning (SAFT) approach that identifies noise sensitive layers in a model, and uses the information to freeze specific layers for noise-injection training. Our results show that SAFT achieves comparable accuracy to noise-injection training and is 2x to 8x faster.

Submitted: Jun 5, 2023