Hardware-Efficient Quantization for Green Custom Foundation Models
Toshiaki Koike-Akino ⋅ Chang Meng ⋅ Volkan Cevher ⋅ Giovanni De Micheli
Abstract
We propose a new hardware-efficient quantization (HEQ) for low-power full-custom foundation models. The HEQ jointly optimizes multiplier hardware and weight quantization to minimize the total power consumption. Exploiting power profile of custom multipliers, our method achieves a significant power reduction up to 20 folds.
Chat is not available.
Successful Page Load