inference/research
verifiedhigh · 0.90updated 2026-10-03

Post Training Quantization of Large Language Models with Microscaling Formats (Tseng et al., 2024)

paper

Fields

id
sources/bench-meth-mxptq
URL
https://arxiv.org/abs/2405.07135
Publisher
arXiv
Kind
paper
Published
2024-05-12
Accessed
2026-10-03
notes
PTQ of LLMs with microscaling formats. Tables 1/2/3 read directly: WikiText-2-test and C4 ppl plus 8-task zero-shot accuracy for INT vs MXINT at fixed calibration, plus the block-size-16 ablation.

Notes

PTQ of LLMs with microscaling formats. Tables 1/2/3 read directly: WikiText-2-test and C4 ppl plus 8-task zero-shot accuracy for INT vs MXINT at fixed calibration, plus the block-size-16 ablation.

Cross-references

This record references no other records.