Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

With Shared Microexponents, A Little Shifting Goes a Long Way

About

This paper introduces Block Data Representations (BDR), a framework for exploring and evaluating a wide spectrum of narrow-precision formats for deep learning. It enables comparison of popular quantization standards, and through BDR, new formats based on shared microexponents (MX) are identified, which outperform other state-of-the-art quantization approaches, including narrow-precision floating-point and block floating-point. MX utilizes multiple levels of quantization scaling with ultra-fine scaling factors based on shared microexponents in the hardware. The effectiveness of MX is demonstrated on real-world models including large-scale generative pretraining and inferencing, and production-scale recommendation systems.

Bita Rouhani, Ritchie Zhao, Venmugil Elango, Rasoul Shafipour, Mathew Hall, Maral Mesmakhosroshahi, Ankit More, Levi Melnick, Maximilian Golub, Girish Varatkar, Lei Shao, Gaurav Kolhe, Dimitry Melts, Jasmine Klar, Renee L'Heureux, Matt Perry, Doug Burger, Eric Chung, Zhaoxia Deng, Sam Naghshineh, Jongsoo Park, Maxim Naumov• 2023

Related benchmarks

TaskDatasetResultRank
Language UnderstandingMMLU
Accuracy65.73
756
Language ModelingWikiText-103
PPL3.58
42
Zero-shot Language ModelingLM Evaluation Harness 0-shot
WG76.4
30
Showing 3 of 3 rows

Other info

Follow for update