// ◈ Quantization, Pruning & Distillation
SmoothQuant
Migrates quantization difficulty from activations to weights for W8A8.
2022 Advanced core
IN PLAIN WORDS
Activations have outliers, weights are smooth. An equivalent per-channel scaling makes both easy to quantize, enabling INT8 weights and activations for faster inference.
Used in: TensorRT-LLM and other engines
01Original paper
SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models
Guangxuan Xiao, Ji Lin, Mickael Seznec, et al. · 2022
arXiv ↗ ✓ Verified · arXiv API
02Where it sits
Learn the left column first; the right column builds on this one.
Learn first
Then you can learn
- Nothing in the atlas builds on it yet
Related:AWQ (Activation-aware Weight Quantization)
Left: prerequisites · right: builds on this · dashed: related. Full map →
03Learning resources
Official docs, explainers, courses and reference code. All links checked.
Reference code
04Where the course teaches it
For AI / data
- Verification: verified, checked 2026-10-21.
- JSON: /api/learn/algorithms.json (id:
smoothquant)