GPTQ
Quantization Method
Overview
Developed byIST Austria researchers
Open source✓ Open Source
Use casePost-training quantization for large language models
Integrates with
Also see
Knowledge graph stats
Claims13
Avg confidence89%
Avg freshness100%
Last updatedUpdated 19 days ago
Trust distribution
100% unverified
GPTQ
concept
Post-training quantization method specifically designed for generative pre-trained transformer models.
Compare with...primary use case
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Post-training quantization for large language models | ○Unverified | High | Fresh | 1 |
supports model
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Large Language Models (LLMs) | ○Unverified | High | Fresh | 1 |
| Transformer models | ○Unverified | High | Fresh | 1 |
published year
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| 2022 | ○Unverified | High | Fresh | 1 |
reduces
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Model memory requirements | ○Unverified | High | Fresh | 1 |
quantization precision
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| 4-bit weights | ○Unverified | High | Fresh | 1 |
based on
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Optimal Brain Quantization (OBQ) method | ○Unverified | High | Fresh | 1 |
integrates with
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| Hugging Face Transformers | ○Unverified | High | Fresh | 1 |
| AutoGPTQ library | ○Unverified | Moderate | Fresh | 1 |
developed by
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| IST Austria researchers | ○Unverified | Moderate | Fresh | 1 |
alternative to
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| AWQ (Activation-aware Weight Quantization) | ○Unverified | Moderate | Fresh | 1 |
| GGML quantization | ○Unverified | Moderate | Fresh | 1 |
open source
| Value | Trust | Confidence | Freshness | Sources |
|---|---|---|---|---|
| true | ○Unverified | Moderate | Fresh | 1 |