// ◈ Quantization, Pruning & Distillation
SparseGPT
One-shot pruning of half a large model's weights without retraining.
SparseGPT 一次性剪枝 2023 Advanced extended
IN PLAIN WORDS
Treats pruning as sparse regression, using approximate second-order information to prune and update weights column-wise. 175B models reach 50% sparsity with little perplexity loss.
Used in: Sparse inference research
01Original paper
SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot
Elias Frantar, Dan Alistarh · 2023
arXiv ↗ ✓ Verified · arXiv API
02Where it sits
Learn the left column first; the right column builds on this one.
Related:BitNet b1.58 (Ternary LLMs)
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:
sparsegpt)