English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Large Language Models (LLMs) are proficient in natural language processingtasks, but their deployment is often restricted by extensive parameter sizesand computational demands. This paper focuses on post-training quantization(PTQ) in LLMs, specifically 4-bit weight and 8-bit activation (W4A8)quantization, to enhance computational efficiency -- a topic less exploredcompared to weight-only quantization. We present two innovative techniques:activation-quantization-aware scaling (AQAS) and sequence-length-awarecalibration (SLAC) to enhance PTQ by considering the combined effects onweights and activations and aligning calibration sequence lengths to targettasks. Moreover, we introduce dINT, a hybrid data format combining integer anddenormal representations, to address the underflow issue in W4A8 quantization,where small values are rounded to zero. Through rigorous evaluations of LLMs,including OPT and LLaMA, we demonstrate that our techniques significantly boosttask accuracies to levels comparable with full-precision models. By developingarithmetic units compatible with dINT, we further confirm that our methodsyield a 2$\times$ hardware efficiency improvement compared to 8-bit integer MACunit.

Enhancing Computation Efficiency in Large Language Models through Weight and Activation Quantization