OmniQuant: Omnidirectionally Calibrated Quantization for Large Language Models (Shi et al.)
paper
Fields
id
sources/qlab-omniquant-paperPublisher
arXiv
Kind
paper
Published
2023-08-25
Accessed
2026-10-03
notes
v3. Learnable Weight Clipping + Learnable Equivalent Transformation; 128 calibration samples on one A100-40G.
Notes
v3. Learnable Weight Clipping + Learnable Equivalent Transformation; 128 calibration samples on one A100-40G.
Cross-references
This record references no other records.