awq-quantization

SKILL.md

4-bit quantization that preserves salient weights based on activation patterns, achieving 3x speedup with minimal accuracy loss.

Timing: 10-15 min for 7B, 1 hour for 70B models.

| Speedup (4-bit) | 2.5-3x | 2x | 1.5x | | Accuracy loss | <5% | 5-10% | 5-15% | | Calibration | Minimal (128-1K tokens) | More extensive | None | | Overfitting risk | Low | Higher | N/A | | Best for | Production inference | GPU inference | Easy integration | | vLLM support | Native | Yes | Limited |

Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner. Source: orchestra-research/ai-research-skills.

View raw