SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models (Xiao et al., ICML 2023)
paper
Fields
id
sources/bench-meth-smoothqPublisher
arXiv / ICML 2023
Kind
paper
Published
2022-11-18
Accessed
2026-10-03
notes
SmoothQuant paper. Tables 3/4/5 read directly: OPT-175B, BLOOM-176B, GLM-130B and OPT-IML-30B W8A8 accuracy and WikiText ppl under the O1/O2/O3 schemes.
Notes
SmoothQuant paper. Tables 3/4/5 read directly: OPT-175B, BLOOM-176B, GLM-130B and OPT-IML-30B W8A8 accuracy and WikiText ppl under the O1/O2/O3 schemes.
Cross-references
This record references no other records.