While research in the field of transformer models has primarily focused on enhancing performance metrics such as accuracy and perplexity, practical applications in industry often necessitate a rigorous consideration of inference latency constraints. Addressing this challenge, we introduce SpeedLimit, a novel Neural Architecture Search (NAS) technique that optimizes accuracy whilst adhering to an upper-bound latency constraint. Our method incorporates 8-bit integer quantization in the search process to outperform the current state-of-the-art technique. Our results underline the feasibility and efficacy of seeking an optimal balance between performance and latency, providing new avenues for deploying state-of-the-art transformer models in latency-sensitive environments.
@article{arxiv.2209.12127,
title = {SpeedLimit: Neural Architecture Search for Quantized Transformer Models},
author = {Yuji Chai and Luke Bailey and Yunho Jin and Matthew Karle and Glenn G. Ko and David Brooks and Gu-Yeon Wei and H. T. Kung},
journal= {arXiv preprint arXiv:2209.12127},
year = {2023}
}