Frantar et al. present GPTQ
AIMade by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made
Presented GPTQ — post-training quantization of a 175B model to 3–4 bits in about 4 GPU hours, fitting it on a single GPU
Why it matters
Standardized quantization as a post-processing step
Source — Primary sources first; secondary reports are labelled “Press”
Paperarxiv.orghttps://arxiv.org/abs/2210.17323Added in Week 35, 2026 · See that week
On the map — ⑲ On-Device line
- Previous stopLLM.int8
- This record · stopGPTQ
- Next stopllama.cpp