# qwopqwop200/GPTQ-for-LLaMa

4 bits quantization of LLaMA using GPTQ

Repository: https://github.com/qwopqwop200/GPTQ-for-LLaMa
Canonical: https://ross.abutalabs.com/products/gptq-for-llama
Language: Python
License: Apache-2.0
License Family: permissive
Last push: 2024-07-13T04:45:28+00:00

## Health v2 (maintenance only)
Score: 30/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 91
- inputs: {"age_days": 1276, "days_push": 781, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3073, forks 453 (observed 2026-08-28T04:07:41.648654+00:00)

## What it is
A Python library that applies GPTQ 4-bit weight quantization to LLaMA large language models, drastically reducing memory usage and checkpoint size. It is now superseded by AutoGPTQ, which the author recommends instead.

## Use cases
- quantize llama models to 4 bits
- run llama 7b on a smaller gpu
- compress large language model checkpoints
- reduce vram usage for llama inference
- convert llama weights with gptq

## When to choose
- you specifically need the original GPTQ-for-LLaMA implementation or its benchmarks
- you are on Linux and want a simple standalone quantization script for LLaMA

## When to avoid
- you want an actively maintained or feature-rich quantization toolkit - use AutoGPTQ instead
- you are on Windows without WSL2, since Triton support is Linux-only
- you need the fastest possible quantized inference kernels

## Facets
- artifact type: library
- maturity: maintenance
- function: llm-inference, machine-learning, deep-learning
- domain: large-language-models, machine-learning, gpu-computing
- platform: python
- tags: quantization, gptq, llama, 4-bit, model-compression, triton, linux, gpu

## Member repositories
- qwopqwop200/GPTQ-for-LLaMa (main) score 30

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:41.648654+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:27:57.305309+00:00, confidence not recorded.
  - readme: https://github.com/qwopqwop200/GPTQ-for-LLaMa (fetched 2026-08-28T04:07:41.648654+00:00, sha 057de12155c0)
- Data as of 2026-08-30T08:39:29.467469+00:00.